Articles

https://doi.org/10.1038/s41562-021-01215-4

Eye movements during text reading align with the

rate of speech production
Benjamin Gagl 1,2,3 ✉, Klara Gregorova 1, Julius Golch1, Stefan Hawelka 4
, Jona Sassenhagen1,
Alessandro Tavano5, David Poeppel 6,7,8 and Christian J. Fiebach 1,2,9

Across languages, the speech signal is characterized by a predominant modulation of the amplitude spectrum between about
4.3 and 5.5 Hz, reflecting the production and processing of linguistic information chunks (syllables and words) every ~200 ms.
Interestingly, ~200 ms is also the typical duration of eye fixations during reading. Prompted by this observation, we demonstrate
that German readers sample written text at ~5 Hz. A subsequent meta-analysis of 142 studies from 14 languages replicates this
result and shows that sampling frequencies vary across languages between 3.9 Hz and 5.2 Hz. This variation systematically
depends on the complexity of the writing systems (character-based versus alphabetic systems and orthographic transparency).
Finally, we empirically demonstrate a positive correlation between speech spectrum and eye movement sampling in low-skilled
non-native readers, with tentative evidence from post hoc analysis suggesting the same relationship in low-skilled native read-
ers. On the basis of this convergent evidence, we propose that during reading, our brain’s linguistic processing systems imprint
a preferred processing rate—that is, the rate of spoken language production and perception—onto the oculomotor system.

S
peech production and perception form a quasi-rhythmic
information processing cycle1. During spoken communica-
tion, our brain entrains to the frequency structure of the
speech signal2,3, suggesting that the temporal structure of the lin-
guistic stimulus drives neural processes in auditory and language
processing systems4. Across languages, the amplitude modulation
spectrum of natural speech peaks consistently in a frequency range
between 4.3 and 5.5 Hz (refs. 5,6), which reflects the fact that informa-
tive signals (for example, syllables7,8) are processed by the listeners’
brains every ~200 ms (ref. 9). Interestingly—and, we hypothesize,
not accidentally—a typical eye fixation during reading has a very
similar duration: between ~200 ms for orthographically transpar-
ent writing systems such as German or Finnish10,11 and ~250 ms for
character-based systems such as Chinese11,12.
Abundant research has used eye-movement recordings to study
reading at high temporal resolution, exploring, for example, how
reading is influenced by word length, word frequency or word pre-
dictability given a sentence context12. Among various measures that
can be derived from eye-movement recordings, timing measures
such as fixation duration are most frequently examined and consid-
ered precise markers of reading speed13,14. These temporally highly
resolved measurements have so far been analysed only at the level
of individual items—typically words. However, other domains of
cognitive research (such as attention15) demonstrate that eye move-
ments can also be subjected to frequency-based analyses. We here
demonstrate that a frequency-based exploration of how written text
is sampled by the eyes can open up new perspectives onto several
fundamental questions related to the process of reading, includ-
ing whether reading is related to spoken language processing (as
recent investigations of word-per-minute measures suggest16) and
whether the visual system’s sampling of linguistic input differs from
eye movements during non-linguistic tasks or between different
languages or writing systems.
To address these foundational questions, we first used an empiri-
cal dataset17 to determine eye movement sampling frequencies for
50 native speakers of German during sentence reading compared
with a non-linguistic control task, using two different methodolo-
gies. Next, to determine the generality of these results and to inves-
tigate possible cross-linguistic differences in the sampling rate of
reading, we conducted a meta-analysis of 124 studies from 14 dif-
ferent languages. To this end, we established a frequency analysis for
fixation durations extracted from published eye-tracking studies.
Finally, we acquired two new datasets, one with 48 non-native and
one with 86 native speakers of German, to directly investigate the
relationship between the sampling frequency of reading and speech
production rates on a subject-by-subject level. The experimental
and meta-analytic results show (1) that written text is sampled in
the same frequency range as spoken language; (2) that the sampling
rate of reading has an upper limit at ~5 Hz, observable in languages
with transparent orthographies; (3) that this rate can be modulated
depending on the complexity of the writing system (for example,
in character-based as opposed to alphabetic scripts or in alphabetic
scripts with opaque grapheme–phoneme mapping); and (4) that a
direct coupling between reading and speech rates is found only in
persons with lower levels of reading skill.
Results
Estimating the sampling rate of reading. Fifty healthy volun-
teers read sentences from the Potsdam Sentence Corpus (144
sentences presented as a whole; 1,138 words in total10) while move-
ments of their right eye were tracked (resolution, 1,000 Hz). As a
non-linguistic control task, the participants scanned ‘z-strings’

1
Department of Psychology, Goethe University Frankfurt, Frankfurt am Main, Germany. 2Center for Individual Development and Adaptive Education
of Children at Risk (IDeA), Frankfurt am Main, Germany. 3Department of Linguistics, University of Vienna, Vienna, Austria. 4Centre for Cognitive
Neuroscience, University of Salzburg, Salzburg, Austria. 5Max Planck Institute for Empirical Aesthetics, Frankfurt am Main, Germany. 6Ernst Struengmann
Institute for Neuroscience, Frankfurt am Main, Germany. 7Department of Psychology, New York University, New York, NY, USA. 8Max-Planck-NYU
Center for Language, Music, and Emotion (CLaME), Frankfurt am Main, Germany. 9Brain Imaging Center, Goethe University Frankfurt,Frankfurt am Main,
Germany. ✉e-mail: benjamin.gagl@univie.ac.at

Nature Human Behaviour | VOL 6 | March 2022 | 429–442 | www.nature.com/nathumbehav 429

Content courtesy of Springer Nature, terms of use apply. Rights reserved
Articles
that were constructed by replacing all letters of the sentence with
the letter ‘z’ (for example, ‘Ein berühmter Maler hat sich selbst
ein Ohr abgeschnitten’ / ‘A famous painter cut off his own ear’
was transformed to ‘Zzz zzzzzzzzz Zzzzz zzz zzzz zzzzzz zzz Zzz
zzzzzzzzzzzzz’; see Methods for the details and ref. 17 for previous
results from this dataset). Given that fixation numbers do not differ
significantly between sentences and z-strings17–19, similar scan paths
are assumed, which qualifies z-strings as valid control stimuli for
reading experiments (Please find multiple analyses that account for
potential differences related to scan path characteristics in results
section below).
Fixation durations. After preprocessing (leading to the removal of
3.1% of the data), we estimated mean fixation durations separately
for each participant and experimental condition. Figure 1a shows
that fixation durations (presented here as subject-specific means)
are shorter for reading than for scanning (average, 197 ms versus
249 ms, respectively; t(49) = 11.1; P < 0.001; Cohen’s d = 1.25; 95%
confidence interval (CI), 0.9 to 1.6). This has been reported previ-
ously for this dataset17 and replicates earlier results for German18,
English20 and French19 in which fixation durations increased from
reading to scanning by 38–42 ms.
Saccade probability. As a first characterization of rhythmic
eye-movement patterns during reading, we plotted the probabil-
ity that a saccade occurs for each sample point after stimulus onset
(Fig. 1b). This analysis demonstrates distinct peaks visible at regular
intervals, providing evidence that eye movements follow a rhythmic
structure in both reading and scanning. Importantly, this rhythmic
pattern is more pronounced and faster during reading. Dominant
sampling rates were estimated directly from fixation durations, as
well as using classical frequency analysis. While the former approach
is important because fixation durations are also the basis for the
subsequent meta-analysis, the latter approach allows us to evaluate
the validity of fixation-duration-based frequency estimation.
Sampling rates. To estimate sampling rates from fixation durations,
we first estimated sampling periods T (that is, the time from the
start of a saccade to the start of the next) by adding to each fixa-
tion duration (n = 112,547) the duration of the preceding saccade.
Figure 1c shows the distribution of all sampling periods across par-
ticipants, separately for reading and z-string scanning. Note that
due to the ex-Gaussian distribution of fixation durations typical for
fixation duration data21, the mean (dashed line) overestimates the
central tendency, whereas the mode (solid line) by definition is a
better representation of the predominant sampling period (Fig. 1c).
Next, we estimated an eye movement sampling frequency f for each
participant and condition, by dividing 1 s by the subject-specific
mode of the sampling period in seconds. This revealed a higher
average sampling rate for reading (5.0 Hz) than for the control
task (4.2 Hz; Fig. 1d). This difference was significant (t(49) = −8.2;
P < 0.001; Cohen’s d = −1.27; 95% CI, −1.7 to −0.9), and 45 of 50
participants showed a numeric reduction in sampling frequency
from reading to scanning (grey lines in Fig. 1d). We find virtually
the same pattern of effects when regressive saccades are removed
(that is, when analysing only single fixation cases; 4.9 Hz and 4.2 Hz
for reading and scanning, respectively; t(49) = −6.9; P < 0.001;
Cohen’s d = −1.13; 95% CI, −1.6 to −0.7) and only slightly higher
values when restricting the analyses to inter-word re-fixations (that
is, fixations after regressive saccades; 5.2 Hz versus 4.6 Hz, respec-
tively; t(48) = −5.5; P < 0.001; Cohen’s d = −0.8; 95% CI, −1.1 to
−0.4; note that one participant was excluded due to the absence of
regressive saccades in the scanning task). Sampling rates of reading
and scanning are thus highly similar between forward-oriented and
regressive eye movements (t(96) = 6.7; P < 0.001; r = 0.6; 95% CI, 0.4
to 0.7). Therefore, all further analyses do not differentiate between
430 Nature Human Behaviour | VOL 6 | March 2022 | 429–442 | www.nature.com/nathumbehav
Content courtesy of Springer Nature, terms of use apply. Rights reserved
Nature Human Behaviour
these cases. Note that estimating sampling rates from the mean
(rather than the mode) of fixation durations results in lower rates
for reading (4.5 Hz) and scanning (3.7 Hz). This results from an
overestimation of the central tendency by the mean in right-skewed
distributions (Fig. 1c) and indicates that this procedure would
be inadequate.
Finally, power spectra of reading versus z-string scanning were
estimated using canonical frequency analysis. For each task, we cre-
ated a time series (resolution, 1,000 Hz) starting with the first sac-
cade of the first participant and ending with the last fixation of the
last participant, with a ‘1’ at the exact time of saccade onset and a
‘0’ elsewhere. Note that saccade onsets are the appropriate event for
generating this time series, as they are the re-occurring event and can
be measured with high accuracy22. Subsequently, the power spectra
of these task-specific event time courses were estimated via Fourier
transform to visualize periodic signal components across subjects
(Methods). Corroborating the results of the first analysis approach,
a prominent peak was found at 5 Hz for reading and a somewhat less
pronounced peak at ~4 Hz for scanning (Fig. 1e). To compare these
estimates between reading and scanning, we next estimated separate
power spectra for each participant. Individual peaks were retrieved,
averaged (Fig. 1f) and statistically compared. This analysis repro-
duces the sampling frequencies estimated from the mode of fixation
durations, with frequencies of 5.0 Hz and 4.4 Hz for reading and
scanning, respectively (t(49) = −7.9; P < 0.001; Cohen’s d = −0.9;
95% CI, −1.1 to −0.6). There was a high correlation between the two
analysis approaches (reading: t(48) = 9.3; P < 0.001; r = 0.80; 95%
CI, 0.7 to 0.9; scanning: t(48) = 5.5; P < 0.001; r = 0.62; 95% CI, 0.4
to 0.8), which underscores the validity of sampling-duration-based
frequency estimations.
To summarize, a quantitative frequency-domain characteriza-
tion of eye-tracking data shows that the predominant sampling
frequency during reading in German, across participants, is ~5 Hz.
This frequency representation of the reading process falls squarely
within the boundaries of the predominant modulation frequencies
of 4.3–5.5 Hz determined for speech signals across languages5,6,
which in turn have a clear reflection in the neuronal response to
speech3. We observed the ~5 Hz peak during reading using two dif-
ferent analysis strategies—that is, when estimating sampling fre-
quencies from saccade and fixation durations and when analysing
the sequence of saccade events in the frequency domain. Attentive
scanning of z-strings shows highly similar scan-path character-
istics compared to reading18,19 but a significantly lower sampling
frequency at ~4.2 Hz, convergent with findings from non-linguistic
attentional reorienting tasks15,23.
An analysis of the pupil response in this same dataset had previ-
ously indicated higher cognitive effort during reading than during
z-string scanning17. This finding most likely reflects the additional
involvement of reading-specific and linguistic processes, such as
lexical–semantic access, beyond the oculomotor sampling itself.
The specific sampling rate observed for reading is thus unlikely to
be driven exclusively by (perceptual or cognitive) features of the
stimulus. In light of the overlap with the rate of spoken language, we
tentatively propose that the observed sampling rate of ~5 Hz may
reflect functional constraints imposed by the interfaced nature that
the process of reading has between visual and linguistic processing
(which developed primarily on the basis of spoken language). We
speculate that the brain’s language systems impose the cortical rate
at which speech is produced and perceived onto oculomotor pro-
gramming systems exclusively during reading, possibly to optimize
language-related information processing.
This hypothesis predicts that the overlap of reading and speech
rates should generalize across languages and writing systems.
However, writing systems differ substantially between languages11,24,
and even within writing systems, the mapping from orthography to
meaning differs between languages25. For example, the letter a in
Nature Human Behaviour Articles
a 400
b
0.0100

Fixation duration (ms)

300

Saccade probability
0.0075

200
0.0050 Reading

Scanning
100 0.0025

0 0
Reading Scanning 0 200 400 600 800 1,000
Task Time relative to the first saccade
of a sentence or z-string (ms)

c d
* Mode Fixation-duration-based
0.006 + Mean estimation

* 6

Individual eye movement

sampling frequency (Hz)
0.004
+
Density

Reading
4
Scanning
0.002
2

0 0
0 200 400 600 800 Reading Scanning
Saccade + fixation durations (ms) Task

e Reading Scanning f Power-spectral-based

estimation
Individual eye movement
sampling frequency (Hz)

6
Power spectral

1 × 10–5
density (a.u.)

5 × 10–6 2

0
1 2 3 4 5 6 7 8 9 10 11 Reading Scanning
Frequency (Hz) Task

Fig. 1 | Reading-related sampling rates. a, Subject-specific mean fixation durations from 50 participants (dots), the overall means (circle) and CIs
(coloured bars) while reading sentences on the Potsdam sentence corpus10 and scanning z-strings. The lines connect reading with z-string scanning data,
per subject, to visualize effects at the single-subject level. The violin plots show the distributions of individual means (blue indicates scanning, and green
indicates reading; similar in d and f). b, Mean saccade probability (across all participants and stimuli, separated by task) relative to the first saccade of
the sentence, with a nonlinear regression line. c, The sampling period T of one event was defined as the duration of a fixation plus its preceding saccade.
Displayed is the distribution of these sampling periods for sentence reading (green) and z-string scanning (blue), with estimated means (plus symbols and
dashed lines) and modes (asterisks and solid lines). d, Subject-specific mean sampling frequencies f (that is, equal to 1/T) and the overall means (crossed
circles) based on the sampling periods shown in c. e, Power spectra for reading and z-string scanning, estimated across all participants using Fourier
transform analysis. f, Individual peak frequencies estimated from individual power spectra and their means (crossed circles). See Methods for details.

cat versus ball maps onto two different speech sounds in English, Cross-linguistic meta-analysis of reading rates. To investigate
whereas it maps onto the same sound in the German translations the language generality of the alignment between speech and read-
of these words (Katze versus Ball). This letter-to-sound correspon- ing rates, we conducted a meta-analysis of sampling frequencies
dence strongly influences reading acquisition26, so that among the during reading in 14 different languages, based on 1,420 fixation
alphabetic writing systems, opaque orthographies (writing systems duration estimates extracted from 124 original studies published
like English with inconsistent letter-to-sound correspondences) are between 2006 and 2016 (see Methods for the selection criteria). In
associated with lower reading accuracy during the first years of learn- addition to this cross-linguistic comparison, we examined possible
ing to read. These differences would be suggestive of cross-linguistic differences between character-based and alphabetic writing sys-
differences in the frequency at which written text can be sampled, tems, as well as the effect of letter-to-sound correspondence among
and recent experimental evidence (such as the observation of longer alphabetic writing systems. We also explored the cross-linguistic
fixation durations for Chinese than for Finnish or English11) seems correlation between eye movement sampling frequencies and
to support this prediction. language-specific peaks of the speech modulation spectra, and the

Nature Human Behaviour | VOL 6 | March 2022 | 429–442 | www.nature.com/nathumbehav 431

Content courtesy of Springer Nature, terms of use apply. Rights reserved
Articles Nature Human Behaviour

400

300
Fixation duration (ms)

200

100

All Arabic Chinese French Polish Italian English Spanish Thai Japanese Hebrew Korean Finnish German Dutch

+ 1 s.d. range of speech rate

7
Eye movement sampling frequency (Hz)

6
Mean range of speech rate

All Arabic Chinese French Polish Italian English Spanish Thai Hebrew Japanese Korean Finnish German Dutch

Fig. 2 | Meta-analysis of reading-related sampling rates. Fixation durations (top) and corresponding eye movement sampling frequencies (bottom) for
14 different languages. The violin plots (left) represent the respective distributions of all 1,420 duration or frequency values extracted from the included
studies, independent of language. The bars reflect CIs, and the circles reflect the means. In the right panel, each dot reflects one study (mean number of
fixation durations per study, 12.4); the bars reflect CIs, and the circles reflect the mean across studies for each language. In the lower panel, the dashed
lines represent the range of the means of the peak amplitude modulation spectrum that was empirically determined for speech in different languages in
independent work5,6. The dotted lines represent the range between the lowest mean minus one standard deviation and the highest mean plus one standard
deviation for the same data (which was manually read out from figure in ref. 5 and from figure 7 in ref. 6). For Arabic, 1 study and 12 fixation durations
are available; for Chinese, 20 and 205; for Dutch, 5 and 45; for English, 65 and 965; for Finnish, 3 and 21; for French, 2 and 3; for German, 14 and 48; for
Hebrew, 3 and 28; for Italian, 1 and 1; for Japanese, 2 and 12; for Korean, 2 and 39, for Polish, 1 and 1; for Spanish, 4 and 10; and for Thai, 3 and 30.

association between reading rates and information density (linguis-
tic information per syllable27) across languages.
All studies selected for inclusion reported mean fixation dura-
tions. However, as discussed above, mean fixation durations are
not a valid representation of the predominant sampling duration in
fixation data and accordingly not the preferred basis for calculating
the sampling rate of reading. We used 29 full empirical datasets to
develop a transformation function that allowed us to estimate the
mode from the mean fixation durations reported in the original
publications. In brief, this involved fitting ex-Gaussian distributions
to the empirical distributions of these datasets, retrieving distri-
butional parameters (including mean and mode) and on this basis
optimizing a regression-based transformation that estimates the
mode from the mean (see Methods and Supplementary Methods for
the details). For the meta-analysis, mean durations were extracted
from published studies and transformed to the mode. The sampling
period T (the interval from saccade onset to the end of the following
fixation; see above) was obtained by adding an estimate (the mode
saccade duration from Study 1; 29 ms) to the mean fixation dura-
tion. Finally, the sampling frequency was calculated as f = 1/T.
Fixation durations and sampling frequency: descriptive statistics.
Figure 2 shows that the majority of mean fixation durations derived
from the reading studies were between 200 and 300 ms (upper
panel), which transforms to mean sampling frequencies between
3.9 and 5.2 Hz (lower panel). Note that languages with only one
original study (Arabic, Italian and Polish) were excluded from
descriptive and further statistical analysis. As expected, the major-
ity of studies were conducted in English28. Ten of the 14 languages
in our meta-analysis fall between the minimum (4.3 Hz) and the
maximum (5.5 Hz) of previously reported5,6 language-specific peaks
of the speech amplitude modulation spectra (Fig. 2, lower panel,
dashed lines). The remaining four languages fell within the range of
one standard deviation around the language-specific speech peaks
(Fig. 2, lower panel, dotted lines). Considering the language-specific
CIs, only for Chinese can we be confident that the sampling rate

432 Nature Human Behaviour | VOL 6 | March 2022 | 429–442 | www.nature.com/nathumbehav

Content courtesy of Springer Nature, terms of use apply. Rights reserved

Nature Human Behaviour
a ference, −0.70 Hz; s.e. = 0.13; see Methods for the details on linear
6 mixed-effects modelling).
sampling frequency (Hz)
4 graphic difficulty of writing systems influences the speed of sam-
Eye movement
2
0
Alphabet Character
b and Dutch (with n = 965, 3, 48 and 45 data points, respectively;
6
sampling frequency (Hz)
4
Eye movement
Dutch
2
English
French
German
0 movement sampling rates (Methods). To control for the strong
100 150 200 250 300
Number of grapheme-to-phoneme rules
Fig. 3 | Comparison of writing systems. a, Character versus alphabetic
contrast, including 205 fixation durations from 20 Chinese reading
studies (orange) and 1,215 fixation durations from 97 studies of reading
in alphabetic languages (grey). b, The effect of language transparency/
opacity. Only studies from alphabetic languages for which the number
of grapheme-to-phoneme rules could be quantitatively estimated from
published computational models (Methods) were used (four languages
with a total of n = 1,025 fixation durations). The dots reflect each study,
and the crossed circles reflect the mean across studies for each language.
The dashed line in b represents the approximation of the language
transparency/opacity effect based on a linear regression.
is lower than the range of the speech amplitude modulation spec-
tra5,6. Of the 1,420 individual sampling rate values derived from the
studies included in our meta-analysis, only 3.0% fell below and only
0.3% were above the range of one standard deviation of the mean
(violin plot in the lower panel of Fig. 2). Nevertheless, the mean
sampling rate of reading observed when averaging across all lan-
guages is at the lower bound of the speech modulation range—that
is, at 4.3 Hz (Fig. 2, lower panel).
Effect of writing system on sampling frequency. The observed
cross-linguistic differences are arguably related to different language
characteristics. One plausible hypothesis is that the higher percep-
tual complexity of character-based scripts (as opposed to alphabetic
scripts24) may modulate the rate at which written text is sampled.
Figure 3a shows that the eye movement sampling frequency is sig-
nificantly lower for Chinese (the only character-based language
included; n = 205 estimated sampling rates from 20 studies; mean,
3.9 Hz) than for alphabetic languages (n = 1,215 sampling rates from
97 studies; mean, 4.5 Hz; t = −5.2; effect size estimate (Est) of dif-
Nature Human Behaviour | VOL 6 | March 2022 | 429–442 | www.nature.com/nathumbehav
Content courtesy of Springer Nature, terms of use apply. Rights reserved
Articles
Effect of orthographic complexity on sampling frequency. Within
alphabetic languages, one plausible hypothesis is that the ortho-
pling the visual input29. To examine this, we quantified orthographic
difficulty as a continuous predictor representing the number of
grapheme-to-phoneme conversion rules30 as defined by compu-
tationally implemented dual-route models of visual word recog-
nition31. Graphemes are letters or letter combinations that map
onto one or multiple speech sounds (phonemes); for illustration,
remember the above example of mapping the grapheme a onto one
(Katze/Ball) versus two (cat/ball) phonemes, requiring one ver-
sus two rules. To date, computational implementations are avail-
able for only five of the nine alphabetic languages included in this
meta-analysis, which restricts this test to English, French, German
Italian, the fifth language, was excluded due to the lack of suffi-
cient data points). A detailed comparison of these language-specific
model implementations can be found in ref. 25. Figure 3b dem-
onstrates that less transparent writing systems (operationalized
as more grapheme-to-phoneme rules) elicit significantly lower
sampling frequencies (t = 3.0; Est = −0.10 Hz; s.e. = 0.03). Highly
transparent orthographies (such as German and Dutch) produce
relatively fast sampling rates around 5 Hz (Fig. 3b).
Effects of speech rate and information density. Lastly, we explored
the effect of cross-linguistic differences in speech rate5,6 and infor-
mation density (information per syllable27) on the observed eye
350 effects of orthographic differences on sampling rates (see above),
linear mixed models were calculated that also included the fac-
tor alphabetic versus character-based script (t > 4; Est < −0.57 Hz;
s.e. < 0.15). Neither the between-language differences in speech fre-
quencies (t = 0.8; Est = −0.03; s.e. = 0.03) nor those in information
density (t = 0.7; Est = −0.03; s.e. = 0.05) showed significant effects
on the eye movement sampling rate (all analyses included Chinese,
Dutch, English, French, Japanese and Spanish).
We further investigated the relationship between speech and
reading rates within alphabetic languages for which estimates of
orthographic complexity (grapheme-to-phoneme rules) could be
taken into account (English, French and Dutch). This analysis also
failed to produce significant effects (t = 1.0; Est = 0.06; s.e. = 0.06).
Still, the result indicated a positive relationship between peak
speech modulation rate and eye movement sampling rate. Note that
we report this analysis despite its low statistical power (with only
three languages) to motivate future investigations of the relation-
ship between speech and reading rates.
The meta-analysis (1) replicates the results obtained for German
in the first section, (2) shows that the eye movement sampling fre-
quencies of most languages fall into the range of the peaks of speech
amplitude modulation spectra (4.3–5.5 Hz)5,6 determined in inde-
pendent research and (3) shows a systematic modulation of read-
ing rates by the perceptual difficulty of orthographic systems. We
found similar average sampling rates for languages of comparable
orthographic transparency levels (for example, German and Dutch)
and the highest reading rates (~5 Hz) in transparent (that is, rela-
tively easy-to-process) writing systems. Our tentative proposal that
the linguistic processing systems underlying speech production and
comprehension provide the temporal frame that ‘drives’ the oculo-
motor machinery during reading would predict a direct relation-
ship between the rate of speech production and the sampling rate
of reading. In the present meta-analysis, this proposal could only be
tested cross-linguistically and using highly aggregated data, and we
found no robust support for this proposal. However, these analyses
433
Articles Nature Human Behaviour

included small sample sizes, as they were limited by the number of a

languages included. To investigate this proposal in more detail, we 0
next conducted two studies that examine the existence of associa-

Speech modulation index

tions between speech and reading rates at a subject-by-subject level.

(amplitude)
−5
Association of individual speech and reading rates. We tested
the correlation between peaks in the speech modulation spectra of
individual speakers and their eye movement sampling rates during
−10
reading in two experiments. First, we tested 48 learners of German
(Study 3), as we expected to observe higher variabilities in both
measures in non-native language learners than in native speakers32
and a more direct relationship between speaking and reading (simi- 0.1 1.0 10.0
lar to letter-by-letter reading in beginning readers33). We recorded Rate (Hz)
eye movements from each participant while reading German sen-
tences (implemented analogously to the reading task in Study 1) b
and a speech sample based on a ‘small-talk interview’ (22 questions,
on average 18 min of speech per participant; range, 6 to 28 min).

Individual frequency (Hz)

We also statistically controlled for individual differences in read- 4
ing proficiency by adding a standard measure of reading skill34,35
(Methods). The eye movement sampling frequency was esti-
mated on the basis of the fixation durations (that is, as in Study 1),
2
and the speech modulation spectrum was examined analogous to
previous reports6.
Figure 4a shows the average speech modulation spectrum across
participants (black line), with a peak at 4.2 Hz, and individual spectra 0
from all participants (grey lines). As expected for language learners, Reading Speech
the peak of the spectrum was below that of native speakers (com-
pare Fig. 1) and on the lower border of the cross-linguistic range of c d
mean speech rates5,6 (compare Fig. 2). Nevertheless, all participants 5 5
produced peaks between 3 Hz and 6 Hz (Fig. 4b)—that is, within
Individual eye movement

Individual eye movement

4 4
the previously reported5,6 range of the standard deviations around
frequency (Hz)

the language-specific mean peaks (dotted line in Fig. 4c). The peaks

of individual speech modulation spectra and eye movement sam-
pling frequencies were in a comparable range (Fig. 4b; confirmed
by a significant equivalence test36: t(47) = −1.8, P = 0.04; equiva-
lence bounds, ±0.33) and positively correlated (Fig. 4c; t(45) = 2.1;
P = 0.04; Est = 0.32; s.e. = 0.15). Note that this correlation effect was
estimated while controlling for individual differences in reading
proficiency (Fig. 4d; t(45) = 2.1; P = 0.04; Est = 0.032; s.e. = 0.016) by
calculating a linear model that estimates the individual eye move-
ment sampling rate with speech modulation rate as a predictor.
In a second, preregistered study (Study 4), we assessed the
relationship between speech modulation spectrum peaks and eye
movement sampling frequencies in a group of 86 native speakers
of German (Methods; preregistration: https://osf.io/mjhkz). We
replicated the finding that the peaks of the speech modulation
spectra and eye movement sampling frequencies were in the same
range (Fig. 5a; equivalence tests for left and right eye: t(85) > 3.9;
P < 0.001; equivalence bounds, ±0.33). The preregistered correlation
analysis showed a small positive (albeit not significant) relation-
ship between eye movement sampling and speech modulation rates
(t(175) = 1.8; P = 0.08; Est = 0.07; s.e. = 0.04). For further explo-
ration (that is, non-registered post-hoc analysis), we separately
investigated and compared four subgroups created by a 2 × 2 com-
bination of reading speed and reading accuracy. Specifically, we
implemented a median split based on reading speed measured with
a standardized German reading test (the adult version of the SLS35;
fast versus slow: median, 78% versus 60%) and, orthogonal to this,
divided the sample on the basis of their sentence comprehension
accuracies in the eye-tracking experiment (errors present versus
absent: median, 0% versus 15%). Only readers with the lowest skill
level (that is, slow and low comprehension performance) showed a
robust positive association (n = 21; Fig. 5b, bottom left; t(37) = 3.4;
P = 0.002; Est = 0.33; s.e. = 0.10). None of the other groups showed
a significant correlation, resulting in a reliable interaction effect
(t(162) = 2.8; P = 0.005; Est = −0.05; s.e. = 0.02), which was also
frequency (Hz)
3 3
2 2
1 1
0 0
3 4 5 20 40 60
Individual speech Individual reading
frequency speed (%
(Hz) sentences read)
Fig. 4 | Relationship of speech and reading rates in non-native German
speakers. a, Speech modulation spectrum from 48 non-native speakers
of German. The y axis shows the speech modulation index6; the x axis
shows the speech modulation rate. For additional comparison, we present
the mean range (dashed lines) and standard deviations (dotted lines) of
the speech amplitude modulation spectra across languages, which were
read out from figure 3c in ref. 5 and from figure 7 in ref. 6. b, Eye movement
sampling frequency in reading and the mean amplitude modulation
spectrum in speech, for each participant. The lines connect the reading
and speech frequencies of each individual, the violin plots represent the
distribution of the data, the bars represent the standard error of the mean
and the circles reflect the mean. c, Positive correlation of the individual peaks
of the speech modulation spectrum (x axis), reflecting each participant’s
speech rate, with the eye movement frequency (y axis) from the same
participants. d, Correlation between the eye movement frequency (y axis)
and a paper-and-pencil-based reading score (x axis) reflecting a positive
association of the eye movement sampling rate and reading performance.
In c and d, we present the individual sampling frequencies corrected for
reading skill and speech frequency, respectively, based on predictions from
the fitted linear regression models used for statistical analysis.
present when the percentage of regressions, skipping and single fix-
ation probabilities (Table 1) were added as covariates to the model.
Note that the low-reading-skill group had a lower reading speed

434 Nature Human Behaviour | VOL 6 | March 2022 | 429–442 | www.nature.com/nathumbehav

Content courtesy of Springer Nature, terms of use apply. Rights reserved

Nature Human Behaviour Articles
a b
Errors No errors
6 6

Individual eye movement frequency (Hz)

4

Fast
Individual frequency (Hz)

2
4

0
6

2
4

Slow
2

0
0
Reading, left eye Reading, right eye Speech 3 4 5 6 3 4 5 6
Individual speech frequency (Hz)

Fig. 5 | Relationship of speech and reading rates in 86 native German speakers. a, Eye movement sampling frequency measured during reading for the
left and right eyes (left and centre plots) and the mean amplitude modulation spectrum of samples of spoken speech (right plot). The grey dots represent
individual data points from all participants; the lines connect the reading and speech frequencies of each participant. The violin plots represent the
distribution of the data, the filled bars represent the standard error of the mean and the circles reflect the mean. b, The correlation between the speech
modulation spectrum (x axis) and the eye movement sampling frequency (y axis). The four panels represent performance subgroups depending on reading
speed (slow versus fast; median split) and whether the participants produced errors in an independent standardized reading test (see Methods for details).

Table 1 | Reading speed, reading comprehension and basic eye tracking measures (fixation durations, skipping probability, single
fixation cases and percentage of regressions) for Study 3 and Study 4
Study 3 Study 4, fast with no Study 4, fast with Study 4, slow with no Study 4, slow with
errors errors errors errors
Reading speed (SLS test; % 26.0 (10.4) 79.0 (6.5) 78.9 (5.3) 58.0 (11.1) 53.6 (9.0)
sentences read)
Reading speed (experiment; in s) 4.8 (2.3) 2.4 (0.7) 2.2 (0.8) 2.4 (0.5) 2.7 (0.6)
Reading comprehension 20 (40) 0 (0) 17 (4.4) 0 (0) 19 (14.5)
(experiment; % errors)
Fixation duration (right eye, in ms) 223 (29) 192 (20) 188 (15) 196 (21) 198 (20)
Fixation duration (left eye, in ms) 225 (30) 192 (20) 189 (15) 197 (21) 199 (20)
Skipping (%) 7.5 (5.8) 19.8 (7.9) 24.0 (10.8) 15.7 (6.9) 14.7 (6.6)
Single fixation cases (%) 35.5 (14.6) 52.3 (9.1) 54.1 (10.2) 56.6 (9.2) 52.7 (9.4)
Regressions (%) 10.9 (4.5) 12.1 (5.3) 9.3 (6.9) 8.6 (4.9) 10.6 (6.0)
For Study 4, the data are presented separately for the four performance-based subgroups. All values reflect means and standard deviations. Statistical comparisons of the four groups from Study 4 (2 × 2,
linear regression models with an error-by-speed-group interaction) yielded a significant main effect of reading speed group (slow versus fast) for SLS-based reading speed (estimate, 0.53; s.e. = 0.22;
t(84) = 2.4; P = 0.017; all other effects, t < 1.7) and an obvious group difference on the percentage of errors (0 versus ~18% errors). For the eye-tracking measures, the only significant effect was a main
effect of reading speed group on the percentage of skipping (estimate, −0.09; s.e. = 0.03; t(84) = 3.4; P < 0.001); no other main effect or interaction effect reached significance (all t < 1.94).

than the slow-only group that produced no errors (t(86) = 18.1;
P < 0.001; Cohen’s d = 2.7; 95% CI, 2.1 to 3.3) but still had a sub-
stantially higher reading speed than the non-native readers from
Study 3 (t(90) = 26.3; P < 0.001; Cohen’s d = 5.5; 95% CI, 4.6 to 6.4;
for general eye-movement characteristics of both experiments,
see Table 1). In sum, we could not find the correlation between
reading and speaking rates in the entire sample, but we found tenta-
tive evidence for the correlation when we restricted the group to
native speakers with low reading skills.
Discussion
This frequency-based investigation of eye movements during read-
ing shows that reading operates in a generally comparable fre-
quency domain as the production and perception of natural speech.
We first reproduced, in a frequency-domain analysis, previous
insights based on fixation duration measures17–19—that is, that eye
movements sample text at a higher rate than during comparable,
cognitively less challenging non-linguistic tasks. More impor-
tantly, we demonstrate that the sampling frequency of reading lies
within the range of previously observed speech rates for one lan-
guage, German. Next, by integrating data from 124 empirical stud-
ies across languages, we show that eye movement sampling varies
between ~3.9 Hz and ~5.2 Hz, indicating a higher variability than
previously assumed. While it was generally believed that average
fixation durations are similar even for very distinct orthographies
such as Chinese and English (Rayner12, p. 1461), our meta-analytic
results show significantly higher sampling rates for alphabetic than
for character-based writing systems. However, average speech rates
have been shown to vary more narrowly around 5 Hz across lan-
guages (that is, 4.3–5.4 Hz in ref. 5 and 4.3–5.5 Hz in ref. 6), a range
that would exclude the lower frequencies we observed for reading.
Our meta-analytic findings indicate that this might result from

Nature Human Behaviour | VOL 6 | March 2022 | 429–442 | www.nature.com/nathumbehav 435

Content courtesy of Springer Nature, terms of use apply. Rights reserved
Articles
differences in the complexity of the underlying orthographies (for
example, character versus alphabetic), such that more computation-
ally ‘difficult’ orthographies might slow down reading relative to
highly transparent alphabetic orthographies.
Subsequently, we demonstrate in two independent empirical
studies that second-language learners (of German) read in a lower
frequency range than native readers (~4.3 Hz versus ~4.7 Hz) and
that only language learners and low-skilled native readers show a
positive correlation between individual reading and speech rates.
Combined, these results suggest that reading (that is, an internally
controlled visual–perceptual process involving sophisticated ocu-
lomotor programming) is remarkably well temporally aligned with
the rate at which spoken language is produced (and perceived). We
tentatively suggest that this observed association between speech
and reading supports the existence of fundamental perceptual prin-
ciples underlying the temporal structure of linguistic information
processing, irrespective of modality16.
Text is a temporally stable visual stimulus. However, our eye
movements impose temporal structure onto the linguistic input
when sequentially sampling a text. The reading process—includ-
ing the oculomotor programmes—thus serves as an interface
between a stable external percept and linguistic processing sys-
tems optimized for analysing sequential speech input. The obser-
vation of faster sampling rates during reading than during parsing
non-linguistic letter strings (Study 1) indicates that sampling rates
are not exclusively driven by the physical layout of the stimulus or
by the cognitive effort of processing the stimulus (in which case
they should have been slower). We tentatively propose that neural
processors dedicated to the linguistic analysis of speech impose
their preferred timing onto the process of reading. Evidence
for the principled possibility of such internally driven entrain-
ment of reading comes from the observation that manipulating
the speed of ‘inner speech’ during reading has a causal effect on
reading speed37–39.
Our meta-analysis demonstrates that the sampling rate of read-
ing varies between languages but falls within the range of speech
rates identified in cross-linguistic studies5,6. The meta-analysis
also shows higher sampling rates for transparent than for opaque
orthographies, which converges with transparency effects within
languages29,40 and cross-linguistic studies investigating reading
development26,41,42. Direct associations between reading and speech
rates could not be established in the meta-analysis given the small
number of languages for which all necessary parameters were
available. Empirical Studies 3 and 4 show this relationship on a
subject-by-subject level, but only in less-skilled readers. This sug-
gests that increasing reading expertise makes the tight control of
reading by linguistic processors in the brain obsolete.
The differential coupling of speech and reading rates in low-skilled
but not high-skilled readers may also result from other phenomena
well-established in reading research, such as word skipping, parafo-
veal preprocessing and re-fixations. Reading is not merely a sequence
of word-to-word fixations. From time to time, we skip words as a
result of parafoveal preprocessing43,44, which describes visual word
recognition based on low-acuity visual information from parafoveal
regions of the retina. Also, words are sometimes fixated multiple
times—for example, to correct perceptual errors after suboptimal
landing at the beginning or end of a word14,18,45 or when semantic
inconsistencies must be resolved by re-reading46,47. We14 and others10
have shown that overall probabilities for word skipping and multiple
fixations on a word are comparable (~20%) when reading the sen-
tence materials used here. However, low-skilled readers show lower
skipping rates (reflecting reduced parafoveal preprocessing48–50) and
more re-fixations on the same word13. Readers with lower skills thus
focus more on fixated words and their components (letters and syl-
lables)51, indicating a greater alignment between the phonological
properties of the words and the eye movements they elicit during
436 Nature Human Behaviour | VOL 6 | March 2022 | 429–442 | www.nature.com/nathumbehav
Content courtesy of Springer Nature, terms of use apply. Rights reserved
Nature Human Behaviour
reading33, which suggests that low-skilled readers sample written
text with a temporal resolution close to the speech processing rate.
In contrast, the faster reading rates of fluent readers indicate that
they utilize the static nature of text better by processing the fix-
ated word and following words (based on parafoveal vision) within
one ‘sample’.
The specific characteristics of fixation behaviour and text pre-
sentation during reading can also provide context for another
intriguing phenomenon—that is, the significantly lower sampling
rates in character-based than in alphabetic writing systems, while
overall reading times for sentences with the same content are com-
parable between the writing systems11. In character-based lan-
guages, fewer fixations per sentence are needed to sample the entire
stimulus, while the increased perceptual complexity and informa-
tion density11 lead to longer fixation durations relative to alphabetic
languages. In the non-linguistic control task, the participants were
presented with stimuli consisting of many repetitions of the same
letter. In this case, information density and perceptual complexity
are low compared with all real-language stimuli. Nevertheless, we
observed longer fixation durations in the z-string scanning task,
which may indicate the presence of qualitatively different cognitive
processes compared with reading.
The frequency representation of reading-related eye-tracking
data that we advance here can be construed as nothing but a
transformation of fixation and saccade duration data. This trans-
formation also comes at the cost of zooming out to a ‘meso-level’
representation of the data1, at which we rely on aggregated data
(that is, one data point per participant), which is against the trend
in eye-movement research of focusing on investigating single words
and using regression methods for detailed analysis of, for example,
the influence of word characteristics52. Still, the frequency perspec-
tive proposed here provides a view of component processes of read-
ing as an interface between linguistic and orthographic processing.
This approach to reading research opens up several interesting new
research questions. For example, it becomes possible to compare
reading behaviour more directly with evidence from other measure-
ment modalities (such as oscillatory brain activation data53,54) and
to other cognitive–psychological domains (such as attention15,55),
which typically do not have the advantage of exact duration mea-
surements for different events of interest (for example, during covert
attention). Maybe most importantly, the frequency perspective on
reading offers direct links to several neurodynamic phenomena in
speech perception5,6, including the observation that dyslexic chil-
dren56,57 and adults58 show altered cortical tracking of speech signals
in the oscillatory domain.
In conclusion, we show that during reading, our eyes ‘sample’
written text in the same frequency range in which speech is pro-
duced and perceived, which suggests that extracting information
from linguistic stimuli follows a similar temporal structure irrespec-
tive of modality. A plausible mechanism is to assume that linguistic
processing has a preferred cortical rate of information uptake and
thus acts as an internal temporal driver for eye movements elic-
ited during reading. Eye movements in reading are thus utilized as
a temporal interface between a stable physical stimulus—written
text—and brain systems that have evolved to process signals whose
temporal structure is constrained by the characteristics of our vocal
tract1. However, our empirical data also demonstrate that a direct
coupling between speech and reading rates is present only in per-
sons with low reading skills, which calls for future work to clarify
the mediating role of reading expertise in the temporal relation-
ship between speech processing and reading rates. We suggest that
the frequency perspective on reading adopted here opens up new
research paths, such as understanding slow or impaired reading,
understanding second language learning or more directly inves-
tigating the commonalities and differences between reading and
other cognitive processes.
Nature Human Behaviour
Methods
Participants in Studies 1, 3 and 4. Fifty (13 male; 18–47 years old; mean = 24
years; students at the University of Salzburg) native speakers of German
participated in Study 1, 49 (13 male; 18–74 years old; mean = 24 years) non-native
German speakers participated in Study 3 and 86 (36 male; 18–53 years old;
mean = 25 years; 5 had to be excluded on the basis of preregistered outlier
correction boundaries for both speech and reading rates; ±3 standard deviations)
German speakers participated in Study 4 after giving informed consent according
to procedures approved by the respective local ethics committee. For Study 1,
see our original publication of this dataset17 for more details. Note that relative
to the previously published report, one participant was added. Participants with
varying mother tongues (Arabic, Azerbaijani, Bulgarian, Chinese, English, Farsi,
French, Georgian, Hungarian, Indonesian, Italian, Japanese, Persian, Russian,
Serbo-Croatian, Spanish, Turkish, Ukrainian, Urdu and Uzbek) for Study 3
and native German participants for Study 4 were recruited on the campus of
Goethe University Frankfurt as part of a more extensive study. Also note that six
participants from Study 3 became literate without the acquisition of an alphabetic
script. The power estimation for Study 4 resulted in a sample size of 90 while
assuming the effect size from Study 3 and a power of 0.9 (see the preregistration for
more details: https://osf.io/mjhkz).
Procedure of Study 1. Movements of the right eye were tracked with a sampling
rate of 1,000 Hz (EyeLink 1000, tower mount system; SR Research). We used a
forehead and chin rest to fixate the head of participants at a distance of 60 cm from
a 21-inch CRT screen. In the reading task, we used the Potsdam Sentence Corpus10,
which consists of 144 sentences and a total of 1,138 words. The participants were
instructed to read silently for comprehension, which was controlled by simple
comprehension questions after 38 of the 144 sentences.
As a non-linguistic control task, the participants performed a z-string scanning
task using stimuli in which all letters of the sentence corpus were replaced by
the letter z (preserving letter case, punctuation and word boundaries). The
participants were instructed to visually scan the meaningless z-strings as if they
were reading; for obvious reasons, no comprehension questions were administered
in this condition. Z-string scanning has been used as a control task in previous
studies17–20. While it is difficult to find a reasonable control task for reading (see, for
example, the Discussion in ref. 18), z-string scanning has proved to be interesting
because participants produce similar scan-path patterns (that is, similar numbers
of fixations) as when reading17–19. Interestingly, while z-string scanning produced
longer mean fixation durations than reading, the pupil response indicated higher
cognitive effort in reading, in the dataset used here17. We take this dissociation
between cognitive effort and reading time as evidence for the operation of
reading-specific cognitive processes that go beyond mere attentional processes.
In each task, a nine-point standard calibration was performed before the ten
practice trials, before the experimental trials and after a break halfway through
the experiment. A calibration was considered accurate when the mean error
was below 0.5° of visual angle. The visual stimuli were presented in black letters
(mono-spaced, bold Courier New font; 14 pt; width, ~0.3°) on a white background
with a 1,024 × 768 pixel resolution and a refresh rate of 120 Hz, using Experiment
Builder software (SR Research). In both tasks, a trial started when an eye fixation
was found at a dot presented 100 pixels from the left margin of the monitor, at the
horizontal level of the fixation cross. For this fixation check, real-time analysis
of eye-tracking data was used to present the sentence only when a fixation of at
least 100 ms was identified on the position of the dot. If no fixation was registered
on the dot for 10 s, a recalibration procedure was initiated. After the fixation
check, the stimulus (that is, a sentence or z-string) appeared, with the centre of
the first word presented at the position of the fixation dot. As a consequence,
the participants always fixated the centre of the first word of the sentence first.
Stimulus presentation was terminated when participants fixated an X in the lower
right corner of the screen after the sentence was read. As noted, in about 25% of
sentences, the presentation was followed by a comprehension question to ensure
that the participants had processed the sentences semantically. This procedure was
practiced in ten trials before the main experiment.
Procedure of Study 3 and Study 4. Eye movement measurements during reading
were acquired using the same stimulus materials and experimental procedures
as in Study 1, with three exceptions: we used a desktop-mount eye tracker and
a horizontal three-point calibration procedure, and we did not implement
the z-string scanning task. All other parameters were unchanged. For Study
4, we measured not only the right eye but also the left eye (that is, binocular
measurement). To acquire a speech sample from each participant, we conducted
a brief interview in German involving 22 questions about, for example, last
weekend’s activities (see the full list of questions in the Supplementary Methods).
Speech was recorded with Audacity software (version 2.1.3; https://www.
audacityteam.org/) on a standard computer.
Data analysis. Fixation durations. The first word of each sentence was excluded
from the analyses, since the first word is known to be contaminated by stimulus
onset effects. A total of 994 words were analysed per subject. For each participant,
all fixation durations from all analysed words were extracted. Words with fixation
Nature Human Behaviour | VOL 6 | March 2022 | 429–442 | www.nature.com/nathumbehav
Content courtesy of Springer Nature, terms of use apply. Rights reserved
Articles
durations shorter than 60 ms and longer than 1,000 ms and saccade durations
longer than 80 ms were removed from the analysis (3.1% of the data) since they
probably reflect machine error. On the basis of the remaining fixation durations,
each participant’s individual mean fixation duration was calculated (separately
for the reading and scanning tasks). To account for the well-known fact that
eye fixation data have an ex-Gaussian distribution (Fig. 1c), the data were
log-transformed, resulting in a normal distribution (Kolmogorov–Smirnov test not
significant; D < 0.14; P > 0.7).
Estimation of the sampling frequency. To estimate the sampling frequency of eye
movements in reading, a repetitive event needed to be identified. We defined the
time between the onset of a saccade and the onset of the following saccade as the
sampling period, which can be transformed into a frequency value. Note that we
used the EyeLink eye tracker’s built-in saccade detection algorithm, which in a
recent comparative evaluation showed the best detection rates for saccade onsets
compared with all other algorithms22. The distribution of the sampling period
is ex-Gaussian, for both reading and z-string scanning (Fig. 1c). Ex-Gaussian
distributions are a convolution of a normal distribution and an exponential
distribution reflecting the rightward skew. As Fig. 1c shows, the central tendency
is best represented by the mode, so that all subsequent calculations were based
on the participant-specific mode of the sampling period (which we denote T).
These subject-specific representations of the predominant sampling period were
transformed to an individual eye movement sampling frequency (f = 1/T).
Power spectrum. We next performed a canonical frequency analysis by estimating
a power spectrum for reading and scanning in Study 1. For Fig. 1e, we estimated
the power spectrum on the basis of a time series starting with the first saccade of
the first participant and ending with the last fixation of the final participant for
each task. For Fig. 1f, the time series was cut into participant-specific time series,
so that individual peaks could be recovered for each participant for each task. The
time series was implemented as a sparse sequence of zeros and ones (resolution,
1,000 entries per second), set to one at time points at which a saccade was initiated,
and zero otherwise. Subsequently, a fast Fourier transform was used to estimate a
power spectrum (power spectral density; psd_welch function from MNE-Python59;
0–100 Hz; length of the fast Fourier transform used, 4,096 samples) for each event
time course separately.
Speech amplitude modulation spectrum in Studies 3 and 4. In the first step, all
non-participant audio signals were removed from the speech samples (that
is, interviewer questions and pauses before answers). To obtain the amplitude
modulation spectrum, we adapted the AM_FM_Spectra script6 (https://github.
com/LeoVarnet/AM_FM_Spectra). The first adaptation divided the recording of
each participant into 10 s speech segments, resulting in a mean of 110 segments
per participant (range, 35 to 167). In Study 4, we used 20 s segments to increase
the efficiency of the analysis (mean number of segments, 55; range, 15 to 81). The
second adaptation was an increase in the resolution of the amplitude modulation
spectrum by decreasing the widths of the modulation filters from three to ten per
octave. After the speech amplitude modulation was estimated for each 10 s speech
segment, we retrieved the frequency at the peak of the modulation spectrum.
Thereafter, we removed outliers by first eliminating unrealistic values lower than
2 and higher than 10 Hz and then removing all values larger and smaller than
two standard deviations from the mean. This procedure removed 3% of the data.
Finally, we estimated the mean across all segments for each participant. Here we
found that one participant, in both studies, had a mean amplitude modulation
spectrum larger than three standard deviations from the mean of the sample; this
participant was excluded from the analysis.
Meta-analysis. We included empirical studies that report eye-tracking results
from natural reading tasks, published between 2006 and 2016. These studies were
identified by the search term eye movement in ‘natural reading’ or ‘sentence reading’
or ‘text reading’ in the PubMed (https://www.ncbi.nlm.nih.gov/pubmed) and
PsychInfo (https://health.ebsco.com/products/psycinfo) databases. Additionally,
ten studies were manually identified (for example, on the basis of reference lists in
the identified papers). From the resulting sample of 124 articles60–183, we extracted
1,420 fixation durations, including mean fixation durations (all fixations on a word
combined; 10% of the meta-analytic dataset), first fixation durations (duration
of the first fixation on a word, which is most often reported in eye-tracking
studies; 67% of the meta-analytic dataset) and single fixation durations (fixation
duration when a word was fixated only once, which is the predominant case for
normal readers10; 23% of the meta-analytic dataset). Note that the results of the
above-reported experiment (Study 1) and its previous analysis17 were not included.
This meta-analytic dataset encompassed 14 languages, with a range from 1 (Arabic,
Italian and Polish) to 65 (English) retrieved papers. Consistent with a general bias
towards English in reading research28, 68% of the fixation durations in our dataset
were from English.
Frequency estimation. To estimate the predominant sampling frequency per
published study, we again have to account for the ex-Gaussian distribution of
fixation duration data. Following the general trend in the eye movement reading
437
Articles
literature, most studies reported only mean fixation durations (except ref. 21, for
which the fitted ex-Gaussian parameters were also reported). For the purposes of
the present meta-analysis, we developed a transformation function that allowed
us to estimate the mode from the mean fixation durations reported in the original
publications. This transformation function was then applied to transform the mean
fixation durations extracted from the published original studies into the mode. In
the final step, the sampling period was determined as the mode fixation duration
plus the mode saccade duration (as estimated from Study 1) and then converted
into a frequency value.
In brief (for the details, see the Supplementary Information), the development
of this function involved fitting ex-Gaussian distributions to the empirical
distributions of fixation durations in 29 original datasets to which we had access
(and which were not included in the meta-analysis; Supplementary Fig. 1a) and
retrieving distributional parameters for each fitted distribution (specifically, μ,
the mean of the normally distributed component; σ, its standard deviation; and
τ, the parameter reflecting the rightward skew, representing the contribution of
the exponential distribution; Supplementary Fig. 1b). This allowed us to apply
a regression analysis to predict the mode (dependent variable) from the mean
fixation duration (predictor variable). Supplementary Fig. 1d presents the final
generalized mean-to-mode transformation function, applied to all possible fixation
durations in the range covered by the meta-analysis. Supplementary Fig. 1e shows
how well the modes of our 29 datasets can be recovered by this function: despite
some unsystematic noise, the numeric transformation was nearly perfect (that
is, t(28) = 4.1; β = 0.95; s.e. = 0.23). We then used the transformation function
to estimate the respective modes from the 1,420 mean fixation durations of the
meta-analysis dataset. To obtain the sampling period T (that is, the interval from the
onset of a saccade until the end of the following fixation; see also Study 1, above), a
saccade duration estimate of 29 ms (that is, the mode of saccade durations from the
reading dataset used in Study 1) was added to each of the mode fixation durations.
This is feasible because saccade durations do not differ much between persons
during reading (see, for example, ref. 184: range, 20–39 ms; mean, 29 ms). Finally, the
sampling period values (T) were transformed into frequency values (f = 1/T).
Writing system comparisons. To explore whether the sampling frequency of reading
is influenced by global characteristics of writing systems and languages, we
implemented four tests. First, reading Chinese (205 fixation duration data points)
was compared with all alphabetic writing systems (1,215 fixation durations). Note
that the Korean (with an alphabetic syllabary orthography185) and Japanese (using
a mixture of Kana and Kanji) studies in the meta-analysis could not be clearly
assigned to the character or alphabet categories and therefore were not included
in this contrast. Second, among the alphabetic scripts, we examined how the
differences in the transparency/opaqueness of the letter-to-sound relationship
influence reading rates using a continuous predictor representing the number of
grapheme-to-phoneme rules as defined by computationally implemented dual-route
models25. A low number of grapheme-to-phoneme rules reflects a high transparency,
meaning that letters more consistently represent only one speech sound. For
example, Italian, Dutch and German are considered transparent orthographies,
with 59, 104 and 130 rules, respectively25. English and French, in contrast, are
typically considered non-transparent (opaque), with 226 and 340 rules, respectively,
because letters regularly map to multiple speech sounds. Third, we investigated the
cross-linguistic relationship between variance in the peak modulation spectra from
speech and mean sampling frequencies in reading. To this end, we retrieved the
modulation spectra for Chinese, Dutch, English, French, Japanese and Spanish from
Fig. 3c in ref. 5 and from Fig. 7 in ref. 6. The modulation spectra varied from 4.3 Hz in
English to 5.4 Hz in Dutch. Finally, in the fourth test, we investigated the relationship
between eye movement sampling in reading and the information density of a
language. This parameter indicates how densely a language codes meaning in texts27.
The density is coded from 0 to 1 and could be retrieved for a subgroup of languages
in the present meta-analysis dataset (Chinese, English, French, German, Italian,
Japanese and Spanish) from ref. 27. Density varied from dense languages such as
Chinese (0.94) to less dense languages such as Japanese (0.49).
All four effects were analysed using linear mixed models186. In addition to the
parameters of interest, we accounted for experimental settings (experiment-based
versus corpus-based studies), for the different eye trackers used (which may also
imply the use of different saccade detection algorithms) and for different fixation
measures reported (mean, single or first fixation duration) by introducing these
parameters into the linear mixed models as fixed effects. For the modulation
spectrum and information density comparisons, we also added a factor contrasting
character-based (that is, Chinese) with alphabetic writing systems to account for
perceptual difficulties, and for all four linear mixed models, we estimated the
random effect on the intercept of the study to account for unspecific differences
between studies. All t values larger than 2 were interpreted as significant187.
Reporting Summary. Further information on research design is available in the
Nature Research Reporting Summary linked to this article.
Data availability
The data for all studies are available at https://osf.io/96vp8/. We have one
restriction—the raw audio of the interviews in Studies 3 and 4. We will not make
438 Nature Human Behaviour | VOL 6 | March 2022 | 429–442 | www.nature.com/nathumbehav
Content courtesy of Springer Nature, terms of use apply. Rights reserved
Nature Human Behaviour
these data publicly available since these interviews contain potentially personal
matters and therefore pose a threat to person-specific data security. However,
the extracted amplitude modulation spectra, which do not include this critical
information, are available.
Code availability
The analysis scripts used for all studies are available at https://osf.io/96vp8/
(https://doi.org/10.17605/OSF.IO/96VP8).
Received: 24 May 2019; Accepted: 10 September 2021;
Published online: 6 December 2021
References
1. Poeppel, D. & Assaneo, M. F. Speech rhythms and their neural foundations.
Nat. Rev. Neurosci. 21, 322–334 (2020).
2. Ding, N., Melloni, L., Zhang, H., Tian, X. & Poeppel, D. Cortical tracking
of hierarchical linguistic structures in connected speech. Nat. Neurosci. 19,
158–164 (2016).
3. Luo, H. & Poeppel, D. Phase patterns of neuronal responses reliably
discriminate speech in human auditory cortex. Neuron 54,
1001–1010 (2007).
4. Martin, A. E. & Doumas, L. A. A. A mechanism for the cortical
computation of hierarchical linguistic structure. PLoS Biol. 15,
e2000663 (2017).
5. Ding, N. et al. Temporal modulations in speech and music. Neurosci.
Biobehav. Rev. https://doi.org/10.1016/j.neubiorev.2017.02.011 (2017).
6. Varnet, L., Ortiz-Barajas, M. C., Erra, R. G., Gervain, J. & Lorenzi, C. A
cross-linguistic study of speech modulation spectra. J. Acoust. Soc. Am. 142,
1976–1989 (2017).
7. Goswami, U. & Leong, V. Speech rhythm and temporal structure:
converging perspectives? Lab. Phonol. 4, 67–92 (2013).
8. Greenberg, S., Carvey, H., Hitchcock, L. & Chang, S. Temporal properties
of spontaneous speech—a syllable-centric perspective. J. Phon. 31, 465–485
(2003).
9. Teng, X., Tian, X., Doelling, K. & Poeppel, D. Theta band oscillations reflect
more than entrainment: behavioral and neural evidence demonstrates an
active chunking process. Eur. J. Neurosci. 48, 2770–2782 (2018).
10. Kliegl, R., Grabner, E., Rolfs, M. & Engbert, R. Length, frequency, and
predictability effects of words on eye movements in reading. Eur. J. Cogn.
Psychol. 16, 262–284 (2004).
11. Liversedge, S. P. et al. Universality in eye movements and reading: a
trilingual investigation. Cognition 147, 1–20 (2016).
12. Rayner, K. Eye movements and attention in reading, scene perception, and
visual search. Q. J. Exp. Psychol. 62, 1457–1506 (2009).
13. Hawelka, S., Gagl, B. & Wimmer, H. A dual-route perspective on eye
movements of dyslexic readers. Cognition 115, 367–379 (2010).
14. Gagl, B., Hawelka, S. & Hutzler, F. A similar correction mechanism in slow
and fluent readers after suboptimal landing positions. Front. Hum. Neurosci.
8, 355 (2014).
15. Hogendoorn, H. Voluntary saccadic eye movements ride the attentional
rhythm. J. Cogn. Neurosci. 28, 1625–1635 (2016).
16. Brysbaert, M. How many words do we read per minute? A review and
meta-analysis of reading rate. J. Mem. Lang. 109, 104047 (2019).
17. Gagl, B., Hawelka, S. & Hutzler, F. Systematic influence of gaze position on
pupil size measurement: analysis and correction. Behav. Res. Methods 43,
1171–1181 (2011).
18. Nuthmann, A., Engbert, R. & Kliegl, R. The IOVP effect in mindless
reading: experiment and modeling. Vis. Res. 47, 990–1002 (2007).
19. Vitu, F., O’Regan, J. K., Inhoff, A. W. & Topolski, R. Mindless reading:
eye-movement characteristics are similar in scanning letter strings and
reading texts. Percept. Psychophys. 57, 352–364 (1995).
20. Rayner, K. & Fischer, M. H. Mindless reading revisited: eye movements
during reading and scanning are different. Percept. Psychophys. 58,
734–747 (1996).
21. Staub, A., White, S. J., Drieghe, D., Hollway, E. C. & Rayner, K.
Distributional effects of word frequency on eye fixation durations.
J. Exp. Psychol. Hum. Percept. Perform. 36, 1280–1293 (2010).
22. Friedman, L., Rigas, I., Abdulin, E. & Komogortsev, O. V. A novel
evaluation of two related and two independent algorithms for eye
movement classification during reading. Behav. Res. Methods 50,
1374–1397 (2018).
23. Landau, A. N. & Fries, P. Attention samples stimuli rhythmically. Curr. Biol.
22, 1000–1004 (2012).
24. Chang, L.-Y., Chen, Y.-C. & Perfetti, C. A. GraphCom: a multidimensional
measure of graphic complexity applied to 131 written languages. Behav. Res.
Methods https://doi.org/10.3758/s13428-017-0881-y (2017).
25. Schmalz, X., Marinus, E., Coltheart, M. & Castles, A. Getting to the bottom
of orthographic depth. Psychon. Bull. Rev. 22, 1614–1629 (2015).
Nature Human Behaviour
26. Seymour, P. H. K., Aro, M. & Erskine, J. M., collaboration with COST
Action A8 network. Foundation literacy acquisition in European
orthographies. Br. J. Psychol. 94, 143–174 (2003).
27. Pellegrino, F., Coupé, C. & Marsico, E. A cross-language perspective on
speech information rate. Language 87, 539–558 (2011).
28. Share, D. L. On the Anglocentricities of current reading research and
practice: the perils of overreliance on an “outlier” orthography. Psychol. Bull.
134, 584–615 (2008).
29. Sereno, S. C. & Rayner, K. Spelling-sound regularity effects on eye fixations
in reading. Percept. Psychophys. 62, 402–409 (2000).
30. Protopapas, A. & Vlahou, E. L. A comparative quantitative analysis
of Greek orthographic transparency. Behav. Res. Methods 41,
991–1008 (2009).
31. Coltheart, M., Rastle, K., Perry, C., Langdon, R. & Ziegler, J. DRC: a dual
route cascaded model of visual word recognition and reading aloud.
Psychol. Rev. 108, 204–256 (2001).
32. Chen, H.-C. Lexical processing in a non-native language: effects of language
proficiency and learning strategy. Mem. Cogn. 18, 279–288 (1990).
33. Gagl, B., Hawelka, S. & Wimmer, H. On sources of the word length effect
in young readers. Sci. Stud. Read. 19, 289–306 (2015).
34. Wimmer, H. Characteristics of developmental dyslexia in a regular writing
system. Appl. Psycholinguist. 14, 1–33 (1993).
35. Auer, M., Guber, G., Wimmer, H. & Mayringer, H. Salzburger
Lese-Screening für die Klassenstufen 1-4 (Hogrefe, 2005).
36. Lakens, D. Equivalence tests: a practical primer for t tests, correlations, and
meta-analyses. Soc. Psychol. Personal. Sci. 8, 355–362 (2017).
37. Alexander, J. D. & Nygaard, L. C. Reading voices and hearing text:
talker-specific auditory imagery in reading. J. Exp. Psychol. Hum. Percept.
Perform. 34, 446–459 (2008).
38. Yao, B. & Scheepers, C. Contextual modulation of reading rate for direct
versus indirect speech quotations. Cognition 121, 447–453 (2011).
39. Stites, M. C., Luke, S. G. & Christianson, K. The psychologist said quickly,
“Dialogue descriptions modulate reading speed!”. Mem. Cogn. 41,
137–151 (2013).
40. Inhoff, A. W. & Topolski, R. Use of phonological codes during eye fixations
in reading and in on-line and delayed naming tasks. J. Mem. Lang. 33,
689–713 (1994).
41. Landerl, K. et al. Phonological awareness and rapid automatized
naming as longitudinal predictors of reading in five alphabetic
orthographies with varying degrees of consistency. Sci. Stud. Read. 23,
220–234 (2019).
42. Moll, K. et al. Cognitive mechanisms underlying reading and spelling
development in five European orthographies. Learn. Instr. 29, 65–77 (2014).
43. Gagl, B., Hawelka, S., Richlan, F., Schuster, S. & Hutzler, F. Parafoveal
preprocessing in reading revisited: evidence from a novel preview
manipulation. J. Exp. Psychol. Learn. Mem. Cogn. 40, 588–595 (2014).
44. Schotter, E. R., Jia, A., Ferreira, V. S. & Rayner, K. Preview benefit in
speaking occurs regardless of preview timing. Psychon. Bull. Rev. 21,
755–762 (2014).
45. Vitu, F., McConkie, G. W., Kerr, P. & O’Regan, J. K. Fixation location effects
on fixation durations during reading: an inverted optimal viewing position
effect. Vis. Res. 41, 3513–3533 (2001).
46. Metzner, P., von der Malsburg, T., Vasishth, S. & Rösler, F. Brain responses
to world knowledge violations: a comparison of stimulus- and
fixation-triggered event-related potentials and neural oscillations. J. Cogn.
Neurosci. 27, 1017–1028 (2015).
47. Schotter, E. R., Tran, R. & Rayner, K. Don’t believe what you read (only
once): comprehension is supported by regressions during reading. Psychol.
Sci. https://doi.org/10.1177/0956797614531148 (2014).
48. Johnson, R. L., Oehrlein, E. C. & Roche, W. L. Predictability and parafoveal
preview effects in the developing reader: evidence from eye movements. J.
Exp. Psychol. Hum. Percept. Perform. 44, 973–991 (2018).
49. Milledge, S. V., Blythe, H. I. & Liversedge, S. P. Parafoveal pre-processing in
children reading English: the importance of external letters. Psychon. Bull.
Rev. https://doi.org/10.3758/s13423-020-01806-8 (2020).
50. Häikiö, T., Bertram, R. & Hyönä, J. Development of parafoveal processing
within and across words in reading: evidence from the boundary paradigm.
Q. J. Exp. Psychol. 63, 1982–1998 (2010).
51. Blythe, H. I. Developmental changes in eye movements and visual
information encoding associated with learning to read. Curr. Dir. Psychol.
Sci. 23, 201–207 (2014).
52. Kliegl, R., Nuthmann, A. & Engbert, R. Tracking the mind during reading:
the influence of past, present, and future words on fixation durations. J.
Exp. Psychol. Gen. 135, 12–35 (2006).
53. Giraud, A.-L. & Poeppel, D. Cortical oscillations and speech processing:
emerging computational principles and operations. Nat. Neurosci. 15,
511–517 (2012).
54. Gross, J. et al. Speech rhythms and multiplexed oscillatory sensory coding
in the human brain. PLoS Biol. 11, e1001752 (2013).
Nature Human Behaviour | VOL 6 | March 2022 | 429–442 | www.nature.com/nathumbehav
Content courtesy of Springer Nature, terms of use apply. Rights reserved
Articles
55. Fiebelkorn, I. C., Saalmann, Y. B. & Kastner, S. Rhythmic sampling within
and between objects despite sustained attention at a cued location. Curr.
Biol. 23, 2553–2558 (2013).
56. Di Liberto, G. M. et al. Atypical cortical entrainment to speech in the right
hemisphere underpins phonemic deficits in dyslexia. NeuroImage 175,
70–79 (2018).
57. Archer, K., Pammer, K. & Vidyasagar, T. R. A temporal sampling basis for
visual processing in developmental dyslexia. Front. Hum. Neurosci. 14,
213 (2020).
58. Leong, V. & Goswami, U. Impaired extraction of speech rhythm from
temporal modulation patterns in speech in developmental dyslexia. Front.
Hum. Neurosci. 8, 96 (2014).
59. Gramfort, A. et al. MNE software for processing MEG and EEG data.
NeuroImage 86, 446–460 (2014).
60. Amenta, S., Marelli, M. & Crepaldi, D. The fruitless effort of growing a
fruitless tree: Early morpho-orthographic and morpho-semantic
effects in sentence reading. J. Exp. Psychol. Learn. Mem. Cogn. 41,
1587–1596 (2015).
61. Angele, B., Laishley, A. E., Rayner, K. & Liversedge, S. P. The effect of
high- and low-frequency previews and sentential fit on word
skipping during reading. J. Exp. Psychol. Learn. Mem. Cogn. 40,
1181–1203 (2014).
62. Ashby, J. Prosody in skilled silent reading: evidence from eye movements.
J. Res. Read. 29, 318–333 (2006).
63. Blythe, H. I., Johnson, R. L., Liversedge, S. P. & Rayner, K. Reading
transposed text: effects of transposed letter distance and consonant-vowel
status on eye movements. Atten. Percept. Psychophys. 76, 2424–2440 (2014).
64. Blythe, H. I., Pagán, A. & Dodd, M. Beyond decoding: phonological
processing during silent reading in beginning readers. J. Exp. Psychol.
Learn. Mem. Cogn. 41, 1244–1252 (2015).
65. Brocher, A., Foraker, S. & Koenig, J.-P. Processing of irregular polysemes in
sentence reading. J. Exp. Psychol. Learn. Mem. Cogn. 42, 1798–1813 (2016).
66. Brzezicka, A., Krejtz, I., von Hecker, U. & Laubrock, J. Eye movement
evidence for defocused attention in dysphoria—a perceptual span analysis.
Int. J. Psychophysiol. 85, 129–133 (2012).
67. Camblin, C. C., Gordon, P. C. & Swaab, T. Y. The interplay of discourse
congruence and lexical association during sentence processing: evidence
from ERPs and eye tracking. J. Mem. Lang. 56, 103–128 (2007).
68. Choi, W., Desai, R. H. & Henderson, J. M. The neural substrates of natural
reading: a comparison of normal and nonword text using eyetracking and
fMRI. Front. Hum. Neurosci. 8, 1024 (2014).
69. Choi, W. & Gordon, P. C. Word skipping during sentence reading: effects of
lexicality on parafoveal processing. Atten. Percept. Psychophys. 76,
201–213 (2014).
70. Choi, W. & Henderson, J. M. Neural correlates of active vision: An fMRI
comparison of natural reading and scene viewing. Neuropsychologia 75,
109–118 (2015).
71. Cop, U., Dirix, N., Drieghe, D. & Duyck, W. Presenting GECO: an
eyetracking corpus of monolingual and bilingual sentence reading.
Behav. Res. Methods https://doi.org/10.3758/s13428-016-0734-0 (2016).
72. Cop, U., Keuleers, E., Drieghe, D. & Duyck, W. Frequency effects in
monolingual and bilingual natural reading. Psychon. Bull. Rev. 22,
1216–1234 (2015).
73. Cui, L. et al. Processing of compound-word characters in reading Chinese:
An eye-movement-contingent display change study. Q. J. Exp. Psychol. 66,
527–547 (2013).
74. Cutter, M. G., Drieghe, D. & Liversedge, S. P. Preview benefit in
English spaced compounds. J. Exp. Psychol. Learn. Mem. Cogn. 40,
1778–1786 (2014).
75. Dambacher, M., Slattery, T. J., Yang, J., Kliegl, R. & Rayner, K. Evidence for
direct control of eye movements during reading. J. Exp. Psychol. Hum.
Percept. Perform. 39, 1468–1484 (2013).
76. Dank, M., Deutsch, A. & Bock, K. Resolving conflicts in natural and
grammatical gender agreement: evidence from eye movements. J.
Psycholinguist. Res. 44, 435–467 (2015).
77. Davis, C. J., Perea, M. & Acha, J. Re(de)fining the orthographic
neighborhood: The role of addition and deletion neighbors in lexical
decision and reading. J. Exp. Psychol. Hum. Percept. Perform. 35,
1550–1570 (2009).
78. Desai, R. H., Choi, W., Lai, V. T. & Henderson, J. M. Toward semantics in
the wild: activation to manipulable nouns in naturalistic reading.
J. Neurosci. 36, 4050–4055 (2016).
79. Deutsch, A., Velan, H. & Michaly, T. Decomposition in a non-concatenated
morphological structure involves more than just the roots: evidence from
fast priming. Q. J. Exp. Psychol. https://doi.org/10.1080/17470218.2016.1250
788 (2016).
80. Dimigen, O., Sommer, W., Hohlfeld, A., Jacobs, A. M. & Kliegl, R.
Coregistration of eye movements and EEG in natural reading: analyses and
review. J. Exp. Psychol. Gen. 140, 552–572 (2011).
439
Articles
81. Dirix, N. & Duyck, W. An eye movement corpus study of the
age-of-acquisition effect. Psychon. Bull. Rev. https://doi.org/10.3758/
s13423-017-1233-8 (2017).
82. Drieghe, D. Foveal processing and word skipping during reading. Psychon.
Bull. Rev. 15, 856–860 (2008).
83. Duñabeitia, J. A., Avilés, A. & Carreiras, M. NoA’s ark: Influence of the
number of associates in visual word recognition. Psychon. Bull. Rev. 15,
1072–1077 (2008).
84. Eiter, B. M. & Inhoff, A. W. Visual word recognition during reading is
followed by subvocal articulation. J. Exp. Psychol. Learn. Mem. Cogn. 36,
457–470 (2010).
85. Farmer, T. A., Yan, S., Bicknell, K. & Tanenhaus, M. K. Form-to-expectation
matching effects on first-pass eye movement measures during reading.
J. Exp. Psychol. Hum. Percept. Perform. 41, 958–976 (2015).
86. Foulsham, T., Farley, J. & Kingstone, A. Mind wandering in sentence
reading: Decoupling the link between mind and eye. Can. J. Exp. Psychol.
Can. Psychol. Expérimentale 67, 51–59 (2013).
87. Frazier, M., Ackerman, L., Baumann, P., Potter, D. & Yoshida, M.
Wh-filler-gap dependency formation guides reflexive antecedent search.
Front. Psychol. 6, 1504 (2015).
88. Frisson, S., Koole, H., Hughes, L., Olson, A. & Wheeldon, L. Competition
between orthographically and phonologically similar words during sentence
reading: evidence from eye movements. J. Mem. Lang. 73, 148–173 (2014).
89. Gagl, B. Blue hypertext is a good design decision: no perceptual
disadvantage in reading and successful highlighting of relevant information.
PeerJ 4, e2467 (2016).
90. Gagl, B., Hawelka, S. & Hutzler, F. A similar correction mechanism in slow
and fluent readers after suboptimal landing positions. Front. Hum. Neurosci.
8, 355 (2014).
91. Gagl, B., Hawelka, S., Richlan, F., Schuster, S. & Hutzler, F. Parafoveal
preprocessing in reading revisited: evidence from a novel preview
manipulation. J. Exp. Psychol. Learn. Mem. Cogn. 40, 588–595 (2014).
92. Gollan, T. H. et al. Frequency drives lexical access in reading but not in
speaking: the frequency-lag hypothesis. J. Exp. Psychol. Gen. 140,
186–209 (2011).
93. Gu, J., Li, X. & Liversedge, S. P. Character order processing in
Chinese reading. J. Exp. Psychol. Hum. Percept. Perform. 41,
127–137 (2015).
94. Hawelka, S., Gagl, B. & Wimmer, H. A dual-route perspective on eye
movements of dyslexic readers. Cognition 115, 367–379 (2010).
95. Henderson, J. M., Choi, W. & Luke, S. G. Morphology of primary visual
cortex predicts individual differences in fixation duration during text
reading. J. Cogn. Neurosci. 26, 2880–2888 (2014).
96. Henderson, J. M., Choi, W., Luke, S. G. & Desai, R. H. Neural correlates of
fixation duration in natural reading: evidence from fixation-related fMRI.
NeuroImage 119, 390–397 (2015).
97. Henderson, J. M. & Luke, S. G. Stable individual differences in saccadic eye
movements during reading, pseudoreading, scene viewing, and scene
search. J. Exp. Psychol. Hum. Percept. Perform. 40, 1390–1400 (2014).
98. Hermena, E. W., Liversedge, S. P. & Drieghe, D. Parafoveal processing of
Arabic diacritical marks. J. Exp. Psychol. Hum. Percept. Perform. 42,
2021–2038 (2016).
99. Howard, P. L., Liversedge, S. P. & Benson, V. Benchmark eye movement
effects during natural reading in autism spectrum disorder. J. Exp. Psychol.
Learn. Mem. Cogn. 43, 109–127 (2017).
100. Huestegge, L., Kunert, H.-J. & Radach, R. Long-term effects of cannabis on
eye movement control in reading. Psychopharmacology (Berl.) 209,
77–84 (2010).
101. Inhoff, A. W., Starr, M. S., Solomon, M. & Placke, L. Eye movements during
the reading of compound words and the influence of lexeme meaning.
Mem. Cognit. 36, 675–687 (2008).
102. Jian, Y.-C. & Ko, H.-W. Investigating the effects of background knowledge
on Chinese word processing during text reading: evidence from eye
movements: eye movements on Chinese word processing. J. Res. Read. 37,
S71–S86 (2014).
103. Jian, Y.-C., Chen, M.-L. & Ko, H. Context effects in processing of Chinese
academic words: an eye-tracking investigation. Read. Res. Q. 48,
403–413 (2013).
104. Jincho, N., Feng, G. & Mazuka, R. Development of text reading in Japanese:
an eye movement study. Read. Writ. 27, 1437–1465 (2014).
105. Johnson, R. L. & Eisler, M. E. The importance of the first and last
letter in words during sentence reading. Acta Psychol. (Amst.) 141,
336–351 (2012).
106. Jordan, T. R., McGowan, V. A., Kurtev, S. & Paterson, K. B. A further look
at postview effects in reading: an eye-movements study of influences from
the left of fixation. J. Exp. Psychol. Learn. Mem. Cogn. 42, 296–307 (2016).
107. Joseph, H. S., Nation, K. & Liversedge, S. P. Using eye movements to
investigate word frequency effects in children’s sentence reading. Sch.
Psychol. Rev. 42, 207 (2013).
440 Nature Human Behaviour | VOL 6 | March 2022 | 429–442 | www.nature.com/nathumbehav
Content courtesy of Springer Nature, terms of use apply. Rights reserved
Nature Human Behaviour
108. Joseph, H. S. S. L. & Liversedge, S. P. Children’s and adults’ on-line
processing of syntactically ambiguous sentences during reading. PLoS ONE
8, e54141 (2013).
109. Joseph, H. S. S. L. et al. Children’s and adults’ processing of anomaly and
implausibility during reading: evidence from eye movements. Q. J. Exp.
Psychol. 61, 708–723 (2008).
110. Juhasz, B. J. & Berkowitz, R. N. Effects of morphological families
on English compound word recognition: a multitask investigation.
Lang. Cogn. Process. 26, 653–682 (2011).
111. Juhasz, B. J. & Rayner, K. The role of age of acquisition and word
frequency in reading: Evidence from eye fixation durations. Vis. Cogn. 13,
846–863 (2006).
112. Kaakinen, J. K. & Hyönä, J. Task effects on eye movements during reading.
J. Exp. Psychol. Learn. Mem. Cogn. 36, 1561–1566 (2010).
113. Kim, Y.-S., Radach, R. & Vorstius, C. Eye movements and parafoveal
processing during reading in Korean. Read. Writ. 25, 1053–1078 (2012).
114. Kliegl, R., Nuthmann, A. & Engbert, R. Tracking the mind during reading:
the influence of past, present, and future words on fixation durations.
J. Exp. Psychol. Gen. 135, 12–35 (2006).
115. Knickerbocker, H., Johnson, R. L. & Altarriba, J. Emotion effects during
reading: Influence of an emotion target word on eye movements and
processing. Cogn. Emot. 29, 784–806 (2015).
116. Koornneef, A. & Vanberkum, J. On the use of verb-based implicit causality
in sentence comprehension: Evidence from self-paced reading and eye
tracking. J. Mem. Lang. 54, 445–465 (2006).
117. Kretzschmar, F., Bornkessel-Schlesewsky, I. & Schlesewsky, M. Parafoveal
versus foveal N400s dissociate spreading activation from contextual fit:.
NeuroReport 20, 1613–1618 (2009).
118. Kuperman, V. & Dyke, Van J. A. Effects of individual differences in verbal
skills on eye-movement patterns during sentence reading. J. Mem. Lang. 65,
42–73 (2011).
119. Li, X., Liu, P. & Rayner, K. Eye movement guidance in Chinese reading: Is
there a preferred viewing location? Vision Res. 51, 1146–1156 (2011).
120. Libben, M. R. & Titone, D. A. Bilingual lexical access in context: evidence
from eye movements during reading. J. Exp. Psychol. Learn. Mem. Cogn. 35,
381–390 (2009).
121. Liu, P. & Li, X. Inserting spaces before and after words affects word
processing differently in Chinese: evidence from eye movements. Br. J.
Psychol. 105, 57–68 (2014).
122. Liversedge, S. P. et al. Universality in eye movements and reading: a
trilingual investigation. Cognition 147, 1–20 (2016).
123. Lou, Y., Liu, Y., Kaakinen, J. K. & Li, X. Using support vector machines to
identify literacy skills: evidence from eye movements. Behav. Res. Methods
https://doi.org/10.3758/s13428-016-0748-7 (2016).
124. Lowder, M. W., Choi, W. & Gordon, P. C. Word recognition during reading:
the interaction between lexical repetition and frequency. Mem. Cognit. 41,
738–751 (2013).
125. Lowder, M. W. & Gordon, P. C. Print exposure modulates the effects of
repetition priming during sentence reading. Psychon. Bull. Rev. https://doi.
org/10.3758/s13423-017-1248-1 (2017).
126. Luo, Y., Yan, M., Yan, S., Zhou, X. & Inhoff, A. W. Syllabic tone articulation
influences the identification and use of words during Chinese sentence
reading: evidence from ERP and eye movement recordings. Cogn. Affect.
Behav. Neurosci. 16, 72–92 (2016).
127. Luo, Y., Yan, M. & Zhou, X. Prosodic boundaries delay the processing of
upcoming lexical information during silent sentence reading. J. Exp.
Psychol. Learn. Mem. Cogn. 39, 915–930 (2013).
128. Ma, G., Li, X. & Pollatsek, A. There is no relationship between preferred
viewing location and word segmentation in Chinese reading. Vis. Cogn. 23,
399–414 (2015).
129. Ma, G., Pollatsek, A., Li, Y. & Li, X. Chinese readers can perceive a word
even when it’s composed of noncontiguous characters. J. Exp. Psychol.
Learn. Mem. Cogn. 43, 158–166 (2017).
130. McDonald, S. A. Effects of number-of-letters on eye movements during
reading are independent from effects of spatial word length. Vis. Cogn. 13,
89–98 (2006).
131. McDonald, S. A., Spitsyna, G., Shillcock, R. C., Wise, R. J. S. & Leff, A. P.
Patients with hemianopic alexia adopt an inefficient eye movement strategy
when reading text. Brain J. Neurol. 129, 158–167 (2006).
132. Metzner, P., von der Malsburg, T., Vasishth, S. & Rösler, F. Brain responses
to world knowledge violations: a comparison of stimulus- and
fixation-triggered event-related potentials and neural oscillations. J. Cogn.
Neurosci. 27, 1017–1028 (2015).
133. Miller, B., Juhasz, B. J. & Rayner, K. The orthographic uniqueness point and
eye movements during reading. Br. J. Psychol. 97, 191–216 (2006).
134. Nikolova, M., Jainta, S., Blythe, H. I. & Liversedge, S. P. Using a dichoptic
moving window presentation technique to investigate binocular
advantages during reading. J. Exp. Psychol. Hum. Percept. Perform. 43,
265–280 (2017).
Nature Human Behaviour
135. Noh, E.-J., Choo, H. & Koh, S. Processing metalinguistic negation: evidence
from eye-tracking experiments. J. Pragmat. 57, 1–18 (2013).
136. Nuthmann, A., Engbert, R. & Kliegl, R. The IOVP effect in mindless
reading: Experiment and modeling. Vision Res. 47, 990–1002 (2007).
137. Nuthmann, A. & Henderson, J. M. Using CRISP to model global
characteristics of fixation durations in scene viewing and reading with a
common mechanism. Vis. Cogn. 20, 457–494 (2012).
138. Pagán, A., Blythe, H. I. & Liversedge, S. P. Parafoveal preprocessing of word
initial trigrams during reading in adults and children. J. Exp. Psychol. Learn.
Mem. Cogn. 42, 411–432 (2016).
139. Perea, M., Giner, L., Marcet, A. & Gomez, P. Does extra interletter
spacing help text reading in skilled adult readers? Span. J. Psychol. 19,
E26 (2016).
140. Philipp, A. M. & Huestegge, L. Language switching between sentences in
reading: Exogenous and endogenous effects on eye movements and
comprehension. Biling. Lang. Cogn. 18, 614–625 (2015).
141. Plummer, P., Perea, M. & Rayner, K. The influence of contextual diversity
on eye movements in reading. J. Exp. Psychol. Learn. Mem. Cogn. 40,
275–283 (2014).
142. Pollatsek, A., Juhasz, B. J., Reichle, E. D., Machacek, D. & Rayner, K.
Immediate and delayed effects of word frequency and word length on eye
movements in reading: a reversed delayed effect of word length. J. Exp.
Psychol. Hum. Percept. Perform. 34, 726–750 (2008).
143. Rau, A. K. et al. Same same, but different: word and sentence reading in
German and English. Sci. Stud. Read. 20, 203–219 (2016).
144. Rayner, K., Li, X., Williams, C. C., Cave, K. R. & Well, A. D. Eye
movements during information processing tasks: individual differences and
cultural effects. Vision Res. 47, 2714–2726 (2007).
145. Rayner, K. & Schotter, E. R. Semantic preview benefit in reading English:
the effect of initial letter capitalization. J. Exp. Psychol. Hum. Percept.
Perform. 40, 1617–1628 (2014).
146. Rayner, K., Slattery, T. J., Drieghe, D. & Liversedge, S. P. Eye movements
and word skipping during reading: effects of word length and predictability.
J. Exp. Psychol. Hum. Percept. Perform. 37, 514–528 (2011).
147. Rayner, K., Yang, J., Schuett, S. & Slattery, T. J. The effect of foveal and
parafoveal masks on the eye movements of older and younger readers.
Psychol. Aging 29, 205–212 (2014).
148. Reali, C., Esaulova, Y., Öttl, A. & von Stockhausen, L. Role descriptions
induce gender mismatch effects in eye movements during reading. Front.
Psychol. 6, 1607 (2015).
149. Reingold, E. M., Reichle, E. D., Glaholt, M. G. & Sheridan, H. Direct lexical
control of eye movements in reading: Evidence from a survival analysis of
fixation durations. Cognit. Psychol. 65, 177–206 (2012).
150. Sainio, M., Hyönä, J., Bingushi, K. & Bertram, R. The role of interword
spacing in reading Japanese: An eye movement study. Vision Res. 47,
2575–2584 (2007).
151. Schad, D. J., Nuthmann, A. & Engbert, R. Eye movements during reading
of randomly shuffled text. Vision Res. 50, 2600–2616 (2010).
152. Schad, D. J., Risse, S., Slattery, T. & Rayner, K. Word frequency in fast
priming: Evidence for immediate cognitive control of eye-movements
during reading. Vis. Cogn. 22, 390–414 (2014).
153. Schotter, E. R., Lee, M., Reiderman, M. & Rayner, K. The effect of
contextual constraint on parafoveal processing in reading. J. Mem. Lang. 83,
118–139 (2015).
154. Staub, A. The effect of lexical predictability on distributions of eye fixation
durations. Psychon. Bull. Rev. 18, 371–376 (2011).
155. Staub, A., White, S. J., Drieghe, D., Hollway, E. C. & Rayner, K.
Distributional effects of word frequency on eye fixation durations. J. Exp.
Psychol. Hum. Percept. Perform. 36, 1280–1293 (2010).
156. Stites, M. C. & Federmeier, K. D. Subsequent to suppression: downstream
comprehension consequences of noun/verb ambiguity in natural reading.
J. Exp. Psychol. Learn. Mem. Cogn. 41, 1497–1515 (2015).
157. Stites, M. C., Federmeier, K. D. & Christianson, K. Do morphemes matter
when reading compound words with transposed letters? evidence from
eye-tracking and event-related potentials. Lang. Cogn. Neurosci. 31,
1299–1319 (2016).
158. Stites, M. C., Federmeier, K. D. & Stine-Morrow, E. A. L. Cross-age
comparisons reveal multiple strategies for lexical ambiguity resolution
during natural reading. J. Exp. Psychol. Learn. Mem. Cogn. 39,
1823–1841 (2013).
159. Titone, D., Libben, M., Mercier, J., Whitford, V. & Pivneva, I. Bilingual
lexical access during L1 sentence reading: the effects of L2 knowledge,
semantic constraint, and L1–L2 intermixing. J. Exp. Psychol. Learn. Mem.
Cogn. 37, 1412–1431 (2011).
160. Tsai, J.-L., Kliegl, R. & Yan, M. Parafoveal semantic information extraction
in traditional Chinese reading. Acta Psychol. (Amst.) 141, 17–23 (2012).
161. Vainio, S., Hyönä, J. & Pajunen, A. Lexical predictability exerts robust
effects on fixation duration, but not on initial landing position during
reading. Exp. Psychol. 56, 66–74 (2009).
Nature Human Behaviour | VOL 6 | March 2022 | 429–442 | www.nature.com/nathumbehav
Content courtesy of Springer Nature, terms of use apply. Rights reserved
Articles
162. Van Assche, E., Drieghe, D., Duyck, W., Welvaert, M. & Hartsuiker, R. J.
The influence of semantic constraints on bilingual word recognition during
sentence reading. J. Mem. Lang. 64, 88–107 (2011).
163. Van der Haegen, L., Cai, Q., Stevens, M. A. & Brysbaert, M.
Interhemispheric communication influences reading behavior. J. Cogn.
Neurosci. 25, 1442–1452 (2013).
164. Velan, H., Deutsch, A. & Frost, R. The flexibility of letter-position flexibility:
Evidence from eye movements in reading Hebrew. J. Exp. Psychol. Hum.
Percept. Perform. 39, 1143–1152 (2013).
165. Veldre, A. & Andrews, S. Parafoveal lexical activation depends on skilled
reading proficiency. J. Exp. Psychol. Learn. Mem. Cogn. 41, 586–595 (2015).
166. Veldre, A. & Andrews, S. Lexical quality and eye movements: Individual
differences in the perceptual span of skilled adult readers. Q. J. Exp. Psychol.
67, 703–727 (2014).
167. Vernet, M. & Kapoula, Z. Binocular motor coordination during saccades and
fixations while reading: a magnitude and time analysis. J. Vis. 9, 2 (2009).
168. Vignali, L., Himmelstoss, N. A., Hawelka, S., Richlan, F. & Hutzler, F.
Oscillatory brain dynamics during sentence reading: a fixation-related
spectral perturbation analysis. Front. Hum. Neurosci. 10, 191 (2016).
169. Inhoff, A. W., Solomon, M. S., Seymour, B. A. & Radach, R. Eye position
changes during reading fixations are spatially selective. Vision Res. 48,
1027–1039 (2008).
170. Wang, A., Zhou, W., Shu, H. & Yan, M. Reading proficiency modulates
parafoveal processing efficiency: evidence from reading Chinese as a second
language. Acta Psychol. (Amst.) 152, 29–33 (2014).
171. Warren, T. & McConnell, K. Investigating effects of selectional restriction
violations and plausibility violation severity on eye-movements in reading.
Psychon. Bull. Rev. 14, 770–775 (2007).
172. White, S. J., Drieghe, D., Liversedge, S. P. & Staub, A. The word frequency
effect during sentence reading: a linear or nonlinear effect of log frequency?
Q. J. Exp. Psychol. https://doi.org/10.1080/17470218.2016.1240813 (2016).
173. White, S. J., Johnson, R. L., Liversedge, S. P. & Rayner, K. Eye movements
when reading transposed text: the importance of word-beginning letters. J.
Exp. Psychol. Hum. Percept. Perform. 34, 1261–1276 (2008).
174. White, S. J., Warrington, K. L., McGowan, V. A. & Paterson, K. B. Eye
movements during reading and topic scanning: effects of word frequency. J.
Exp. Psychol. Hum. Percept. Perform. 41, 233–248 (2015).
175. Winskel, H. Orthographic and phonological parafoveal processing of
consonants, vowels, and tones when reading Thai. Appl. Psycholinguist. 32,
739–759 (2011).
176. Winskel, H., Perea, M. & Ratitamkul, T. On the flexibility of letter position
coding during lexical processing: Evidence from eye movements when
reading Thai. Q. J. Exp. Psychol. 65, 1522–1536 (2012).
177. Winskel, H., Radach, R. & Luksaneeyanawin, S. Eye movements when
reading spaced and unspaced Thai and English: a comparison of Thai–
English bilinguals and English monolinguals. J. Mem. Lang. 61,
339–351 (2009).
178. Wotschack, C. & Kliegl, R. Reading strategy modulates parafoveal-on-foveal
effects in sentence reading. Q. J. Exp. Psychol. 66, 548–562 (2013).
179. Yan, M. & Kliegl, R. CarPrice versus CarpRice: Word boundary ambiguity
influences saccade target selection during the reading of Chinese sentences.
J. Exp. Psychol. Learn. Mem. Cogn. 42, 1832–1838 (2016).
180. Yan, M., Kliegl, R., Richter, E. M., Nuthmann, A. & Shu, H. Flexible
saccade-target selection in Chinese reading. Q. J. Exp. Psychol. 63, 705–725
(2010).
181. Yang, J. Preview effects of plausibility and character order in reading
Chinese transposed words: evidence from eye movements: preview reading
of transposed words. J. Res. Read. 36, S18–S34 (2013).
182. Yen, M.-H., Tsai, J.-L., Tzeng, O. J.-L. & Hung, D. L. Eye movements and
parafoveal word processing in reading Chinese. Mem. Cognit. 36,
1033–1045 (2008).
183. Yen, M.-H., Radach, R., Tzeng, O. J.-L. & Tsai, J.-L. Usage of statistical cues
for word boundary in reading Chinese sentences. Read. Writ. 25,
1007–1029 (2012).
184. Rayner, K. Eye movement latencies for parafoveally presented words. Bull.
Psychon. Soc. 11, 13–16 (1978).
185. Taylor, I. in Processing of Visible Language (eds Kolers, P. A. et al.) 67–82
(Springer, 1980).
186. Bates, D., Mächler, M., Bolker, B. & Walker, S. Fitting linear mixed-effects
models using lme4. J. Stat. Softw. 67, 1–48 (2015).
187. Kliegl, R., Wei, P., Dambacher, M., Yan, M. & Zhou, X. Experimental effects
and individual differences in linear mixed models: estimating the
relationship between spatial, object, and attraction effects in visual
attention. Front. Psychol. 1, 238 (2011).
Acknowledgements
The research leading to these results has received funding from the European
Community’s Seventh Framework Programme (FP7/2013) under grant agreement no.
617891 awarded to C.J.F. and from the European Community’s Horizon 2020 Programme
441
Articles
under grant agreement no. 707932 awarded to B.G. The funders had no role in study
design, data collection and analysis, decision to publish or preparation of the manuscript.
We thank J. Müller for helpful comments on a previous version of the manuscript.
Author contributions
B.G., D.P. and C.J.F. designed the research. B.G. and S.H. performed Study 1. B.G. and
J.G. performed Study 2 (meta-analysis). B.G. and K.G. performed Studies 3 and 4.
B.G., A.T. and J.S. analysed the data. B.G. and C.J.F. wrote the paper. All authors gave
comments on the paper during the process.
Competing interests
The authors declare no competing interests.
Nature Human Behaviour
Additional information
Supplementary information The online version contains supplementary material
available at https://doi.org/10.1038/s41562-021-01215-4.
Correspondence and requests for materials should be addressed to Benjamin Gagl.
Peer review information Nature Human Behaviour thanks Denis Drieghe, Kristen
Pammer and the other, anonymous, reviewer(s) for their contribution to the peer review
of this work. Peer reviewer reports are available.
Reprints and permissions information is available at www.nature.com/reprints.
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in
published maps and institutional affiliations.
© The Author(s), under exclusive licence to Springer Nature Limited 2021

442 Nature Human Behaviour | VOL 6 | March 2022 | 429–442 | www.nature.com/nathumbehav

Content courtesy of Springer Nature, terms of use apply. Rights reserved

 nature portfolio | reporting summary
Corresponding author(s): Benjamin Gagl
Last updated by author(s): Jul 21, 2021

Reporting Summary
Nature Portfolio wishes to improve the reproducibility of the work that we publish. This form provides structure for consistency and transparency
in reporting. For further information on Nature Portfolio policies, see our Editorial Policies and the Editorial Policy Checklist.

Statistics
For all statistical analyses, confirm that the following items are present in the figure legend, table legend, main text, or Methods section.
n/a Confirmed
The exact sample size (n) for each experimental group/condition, given as a discrete number and unit of measurement
A statement on whether measurements were taken from distinct samples or whether the same sample was measured repeatedly
The statistical test(s) used AND whether they are one- or two-sided
Only common tests should be described solely by name; describe more complex techniques in the Methods section.

A description of all covariates tested

A description of any assumptions or corrections, such as tests of normality and adjustment for multiple comparisons
A full description of the statistical parameters including central tendency (e.g. means) or other basic estimates (e.g. regression coefficient)
AND variation (e.g. standard deviation) or associated estimates of uncertainty (e.g. confidence intervals)

For null hypothesis testing, the test statistic (e.g. F, t, r) with confidence intervals, effect sizes, degrees of freedom and P value noted
Give P values as exact values whenever suitable.

For Bayesian analysis, information on the choice of priors and Markov chain Monte Carlo settings
For hierarchical and complex designs, identification of the appropriate level for tests and full reporting of outcomes
Estimates of effect sizes (e.g. Cohen's d, Pearson's r), indicating how they were calculated
Our web collection on statistics for biologists contains articles on many of the points above.

Software and code

Policy information about availability of computer code
Data collection Experiment Builder software (SR Research, Ontario, Canada) for presentation of sentences and z-strings in eyetracking experiments. Speech
was recorded with the Audacity software (Version 2.1.3; https://www.audacityteam.org/). Both were compiled stand allone software running
on Windows PC.

Data analysis We implemented eye-tracking data preprocessing in Perl on a MAC system, speech data preprocessing in MATLAB and the data preprocessing
for the Fourier transform control analysis in Python both on a Ubuntu system. All further processing steps and analysis were implemented in
R. Please find all scripts here https://osf.io/96vp8/.
For manuscripts utilizing custom algorithms or software that are central to the research but not yet described in published literature, software must be made available to editors and
reviewers. We strongly encourage code deposition in a community repository (e.g. GitHub). See the Nature Portfolio guidelines for submitting code & software for further information.

Data
Policy information about availability of data
March 2021

All manuscripts must include a data availability statement. This statement should provide the following information, where applicable:
- Accession codes, unique identifiers, or web links for publicly available datasets
- A description of any restrictions on data availability
- For clinical datasets or third party data, please ensure that the statement adheres to our policy

Provide your data availability statement here.

Content courtesy of Springer Nature, terms of use apply. Rights reserved

 nature portfolio | reporting summary
Field-specific reporting
Please select the one below that is the best fit for your research. If you are not sure, read the appropriate sections before making your selection.
Life sciences Behavioural & social sciences Ecological, evolutionary & environmental sciences
For a reference copy of the document with all sections, see nature.com/documents/nr-reporting-summary-flat.pdf

Behavioural & social sciences study design

All studies must disclose on these points even when the disclosure is negative.
Study description Three quantitative studies (Study 1, 3 & 4) and one meta-analytic study (Study 2) are included.

Research sample Fifty (13 male; 18–47 years old; M=24 years; students at University of Salzburg) native speakers of German participated in Study 1,
forty-nine (13 male; 18–74years old; M=24 years) non-native German speakers participated in Study 3, and eighty-six (36 male; 18–
53years old; M=25years; five had to be excluded based on preregistered outlier correction boundaries for both speech and reading
rates; +-3 standard deviations) German speakers participated in Study 4 after giving informed consent according to procedures
approved by the respective local ethics committee. For Study 1, see Gagl, Hawelka & Hutzler, 2011, BRM for more details. Note that
relative to the previously published report, one participant was added. For Study 3, participants with varying mother tongues (Arabic,
Azerbaijani, Bulgarian, Chinese, English, Farsi, French, Georgian, Indonesian, Italian, Japanese, Persian, Russian, Serbo-Croatian,
Spanish, Turkish, Ukrainian, Hungarian, Urdu, and Uzbek) and, for Study 4, native German participants were recruited on the campus
of Goethe University Frankfurt as part of a more extensive study. Also, note that six participants from Study 3 became literate
without the acquisition of an alphabetic script.

Sampling strategy
Datasets for studies 1 and 3 were available before, and the sample size of 49 and 50 participants typically prevent the inflation of
overestimation of effects due to small sample sizes (e.g., see Cremers HR, Wager TD, Yarkoni T (2017) The relation between statistical
power and inference in fMRI. PLoS ONE 12(11): e0184923. https://doi.org/10.1371/journal.pone.0184923 shows stable effects with
N > 30). For the dataset in study 4, we provide a pre-registration, including a power calculation (https://osf.io/mjhkz). We
implemented a Monte-Carlo simulation procedure (i.e., randomly sampling data and re-estimating the general linear model 10,000
times with different participants). With this procedure, we learned that a total of 90 participants must be collected to gain a power
value of .9. After excluding five participants due to a pre-registered exclusion criterion, data from 86 participants were analyzed. For
study 1, student participants from the University of Salzburg were selected, for study 3 non-native German speakers, and for study 4
native German speakers, both from the Frankfurt area were selected.

Data collection In reading and scanning tasks used in Study 1, 3 & 4 all stimuli were presented on computer screens. The researcher presented the
interview questions, and the reading scores were implemented via a paper-pencil test. One researcher was testing one participant,
and researchers testing were blind to the hypothesis of the current study.

Timing For Study 1, we relied on a dataset first presented in Gagl, Hawelka, & Hutzler 2011. For Study 3, all data was collected in 2018. For
Study 4, all data was collected in 2019 and 2020.

Data exclusions In Study 3 we found that one and study 4 five participant had a mean amplitude modulation spectrum, which was larger than three
standard deviations from the mean of the sample; this participant was excluded from the analysis. See analysis section for details.

Non-participation No participants dropped out.

Randomization No allocation into experimental groups.

Reporting for specific materials, systems and methods

We require information from authors about some types of materials, experimental systems and methods used in many studies. Here, indicate whether each material,
system or method listed is relevant to your study. If you are not sure if a list item applies to your research, read the appropriate section before selecting a response.

Materials & experimental systems Methods

n/a Involved in the study n/a Involved in the study
Antibodies ChIP-seq
Eukaryotic cell lines Flow cytometry
March 2021

Palaeontology and archaeology MRI-based neuroimaging

Animals and other organisms
Human research participants
Clinical data
Dual use research of concern

Content courtesy of Springer Nature, terms of use apply. Rights reserved

Human research participants

nature portfolio | reporting summary

Policy information about studies involving human research participants
Population characteristics See above.

Recruitment In booth studies advertisements were used to recruit participants.

Ethics oversight Study 1 was approved by the ethics board of the University of Salzburg and Studies 3/4 was approved by the ethics board of
the Goethe University of Frankfurt
Note that full information on the approval of the study protocol must also be provided in the manuscript.

March 2021

Content courtesy of Springer Nature, terms of use apply. Rights reserved

Terms and Conditions
Springer Nature journal content, brought to you courtesy of Springer Nature Customer Service Center GmbH (“Springer Nature”).
Springer Nature supports a reasonable amount of sharing of research papers by authors, subscribers and authorised users (“Users”), for small-
scale personal, non-commercial use provided that all copyright, trade and service marks and other proprietary notices are maintained. By
accessing, sharing, receiving or otherwise using the Springer Nature journal content you agree to these terms of use (“Terms”). For these
purposes, Springer Nature considers academic use (by researchers and students) to be non-commercial.
These Terms are supplementary and will apply in addition to any applicable website terms and conditions, a relevant site licence or a personal
subscription. These Terms will prevail over any conflict or ambiguity with regards to the relevant terms, a site licence or a personal subscription
(to the extent of the conflict or ambiguity only). For Creative Commons-licensed articles, the terms of the Creative Commons license used will
apply.
We collect and use personal data to provide access to the Springer Nature journal content. We may also use these personal data internally within
ResearchGate and Springer Nature and as agreed share it, in an anonymised way, for purposes of tracking, analysis and reporting. We will not
otherwise disclose your personal data outside the ResearchGate or the Springer Nature group of companies unless we have your permission as
detailed in the Privacy Policy.
While Users may use the Springer Nature journal content for small scale, personal non-commercial use, it is important to note that Users may
not:

1. use such content for the purpose of providing other users with access on a regular or large scale basis or as a means to circumvent access
control;
2. use such content where to do so would be considered a criminal or statutory offence in any jurisdiction, or gives rise to civil liability, or is
otherwise unlawful;
3. falsely or misleadingly imply or suggest endorsement, approval , sponsorship, or association unless explicitly agreed to by Springer Nature in
writing;
4. use bots or other automated methods to access the content or redirect messages
5. override any security feature or exclusionary protocol; or
6. share the content in order to create substitute for Springer Nature products or services or a systematic database of Springer Nature journal
content.
In line with the restriction against commercial use, Springer Nature does not permit the creation of a product or service that creates revenue,
royalties, rent or income from our content or its inclusion as part of a paid for service or for other commercial gain. Springer Nature journal
content cannot be used for inter-library loans and librarians may not upload Springer Nature journal content on a large scale into their, or any
other, institutional repository.
These terms of use are reviewed regularly and may be amended at any time. Springer Nature is not obligated to publish any information or
content on this website and may remove it or features or functionality at our sole discretion, at any time with or without notice. Springer Nature
may revoke this licence to you at any time and remove access to any copies of the Springer Nature journal content which have been saved.
To the fullest extent permitted by law, Springer Nature makes no warranties, representations or guarantees to Users, either express or implied
with respect to the Springer nature journal content and all parties disclaim and waive any implied warranties or warranties imposed by law,
including merchantability or fitness for any particular purpose.
Please note that these rights do not automatically extend to content, data or other material published by Springer Nature that may be licensed
from third parties.
If you would like to use or distribute our Springer Nature journal content to a wider audience or on a regular basis or in any other manner not
expressly permitted by these Terms, please contact Springer Nature at

onlineservice@springernature.com