You are on page 1of 5

Neuroscience Letters 485 (2010) 138–142

Contents lists available at ScienceDirect

Neuroscience Letters
journal homepage: www.elsevier.com/locate/neulet

Impulse inhibition in people with Internet addiction disorder:


Electrophysiological evidence from a Go/NoGo study
Guangheng Dong a,∗ , Qilin Lu b,c , Hui Zhou a , Xuan Zhao a
a
Department of Psychology, Zhejiang Normal University, PR China
b
Institute of Neuroinformatics, Dalian University of Technology, PR China
c
Courtesy Research Assistant, University of Oregon, United States

a r t i c l e i n f o a b s t r a c t

Article history: We investigated response inhibition in people with Internet addiction disorder (IAD) by recording event-
Received 17 January 2010 related brain potentials during a Go/NoGo task. Twelve IAD-afflicted and 12 normal university students
Received in revised form 2 July 2010 participated in the study. Results show that the IAD group exhibited lower NoGo-N2 amplitude, higher
Accepted 1 September 2010
NoGo-P3 amplitude, and longer NoGo-P3 peak latency than the normal group. The results also suggest
that the IAD students had lower activation in the conflict detection stage than the normal group; thus,
Keywords:
they had to engage in more cognitive endeavors to complete the inhibition task in the late stage. In
Event-related potentials
addition, the IAD students showed less efficiency in information processing and lower impulse control
Go/NoGo
Impulse control ability
than their normal peers.
Internet addiction disorder © 2010 Elsevier Ireland Ltd. All rights reserved.

Internet addiction disorder (IAD), also described as pathological impaired in people with IAD. If IAD is related to lower impulse con-
Internet use, is the inability of an individual to control his or her trol, this research is likely to demonstrate that neuropsychological
use of the Internet, eventually causing psychological, social, and/or characteristics of IAD may be similar to those of other disorders.
work difficulties [11,30]. Studies reveal the correlation between However, few electrophysiological studies about the relationship
low impulse control and other addictive behaviors, such as patho- between IAD and impulse control have been done. Thus, this study
logical gambling, substance abuse, and alcohol abuse. Barnes et al. aims to assess whether Internet addiction is related to impaired
found that lower impulse control is a significant predictor of alco- impulse control based on electrophysiological evidence.
hol misuse for females and delinquency for males [1]. Vitaro et A useful way to assess response inhibition is through Go/NoGo
al. used a prospective-longitudinal design to investigate whether tasks, which always consist of two stimuli: a Go stimulus requiring a
measuring low impulse control in 12–14-year-olds could predict response (usually a button press) and a NoGo stimulus that requires
engagement in gambling in late adolescence [29]. Moeller et al. the inhibition of the response [13,21,23]. To elicit the response
found that impulsivity is a significant predictor of cocaine use impulse, more Go trials were conducted than NoGo trials. Two
and treatment retention [25]. Cavedini et al. also explored the major components of event-related potentials (ERPs) have been
relationship between the ventromedial orbitofrontal circuits and consistently linked with the response inhibition in Go/NoGo tasks.
pathological gambling [10]. The first is an enhanced negative component (NoGo-N2) at approx-
Researchers believe that Internet addiction is an impulse dis- imately 200 ms post-stimulus onset in response to NoGo stimuli,
order or at least related to impulse control disorder [2,30] because and is maximal in the frontal areas [5]. N2 may represent response
pathological gamblers, drug addicts, and alcohol abusers may share inhibition [18] or the process of conflict monitoring [28]. The sec-
similar neuropsychological and personality characteristics with ond major ERP component is an enhanced wave (NoGo-P3) elicited
Internet addicts. Using questionnaires, Cao showed a specific rela- within a 300–500 ms time window [6]. NoGo-P3 shows a fronto-
tionship between impulse control and Internet addiction [9]. In central maximum as opposed to the centro-parietal maximum of
the current study, the neuroactivity of the impulse control pro- Go-P3 [6]. Infrequent stimuli-related P3 is an important compo-
cess in IAD was explored to determine whether impulse control is nent in oddball tasks. NoGo-P3 is also believed to be related to
response inhibition and indexing a late stage of the inhibitory pro-
cess, such as, response evaluation or successful response inhibition
[14,16,31].
∗ Corresponding author at: Department of Psychology, Zhejiang Normal Univer-
With regard to the relation between brain activity and impulse
sity, 688 Yingbin Road, Jinhua City, Zhejiang Province, Postal Code 321004, PR China.
control, Kenemans et al. believed that the direct reflections of brain
Tel.: +86 579 8228 2961; fax: +86 579 8228 2549.
E-mail address: dongguangheng@zjnu.edu.cn (G. Dong). activity suggest that the mechanisms of expectation and atten-

0304-3940/$ – see front matter © 2010 Elsevier Ireland Ltd. All rights reserved.
doi:10.1016/j.neulet.2010.09.002
G. Dong et al. / Neuroscience Letters 485 (2010) 138–142 139

tion govern behavioral manifestations of impulsivity [22]. Studies fier. The electroencephalogram was continuously recorded with
on attention-deficit hyperactivity disorder also reveal that atten- a sample rate of 250 Hz. Whenever possible, impedances were
tion deficiency is an important factor in lower impulse control reduced to less than 5 k before recording with vertical elec-
[3]. IAD-afflicted people show some attention deficiency features trooculograms recorded at the left orbital rim, and horizontal
[15]. Thus, we believe that impulse control in IAD-afflicted indi- electrooculograms recorded at the right orbital rim.
viduals differs from that in normal people and can be detected EEG epochs of 800 ms, including 200 ms of pre-stimulus time as
by psycho-physiological evidence. The present study aims to baseline, were offline averaged only using correct trials according
find a neural index underlying the response inhibition difference to the stimuli (Go and NoGo). Epochs with artifacts exceeding
between Internet-addicted and normal people by using an ERP ±50 ␮V at any electrode were omitted from further analysis.
technique. The mechanism that caused the behavior of the par- The peak latencies were calculated by the software Netstation
ticipants was explored. As discussed, N2 is believed to be related (Electrical Geodesics Inc., Eugene, OR, USA). The N2 amplitude was
to the process of conflict monitoring, and P3 to response eval- calculated at the negative maximum between 180 and 220 ms
uation. These two mental processes are fundamental abilities in and the P3 amplitude was calculated at the positive maximum
the impulse inhibition process. Internet-addicted participants were between 250 and 450 ms. Because the number of trials in Go
expected to show some difference in N2 and P3 compared with their and NoGo conditions was not equal (4:1), we selected only one
normal peers. quarter of the Go trials randomly for further comparison. The
Twenty-four healthy right-handed (determined by digitizing following sites were chosen for statistical analysis: Fz, F1, F2,
tablet method) university students participated in this study. All F3, F4 (5 frontal sites), FCz, FC1, FC2, FC3, FC4 (5 frontal-central
participants were paid 30 dollars for their participation. The IAD sites), Cz, C1, C2, C3, C4 (5 central sites), CPz, CP1, CP2, CP3, CP4 (5
group (12 males, 20.47 ± 4.12 years old) were selected by Young’s central-parietal sites), and Pz, P1, P2, P3, P4 (5 parietal sites). The
Internet addiction test (IAT) [30] in which 8 items were presented, frontal and central sites (15 sites) were selected for comparison
and only those who scored 5 or higher were considered Internet in N2. All of the 25 sites were selected for comparison in P3. ERP
addicts. A strict selection criterion was used: only those who scored amplitudes and latencies were analyzed using repeated measures
7 or higher were allowed to participate. The students from the nor- ANOVA with electrode sites (frontal/central/parietal) × stimulus
mal group (12 males, 20.19 ± 4.47 years old) were recruited from (Go/NoGo) as within-subject factors and group (Internet
the same university and scored lower than 4 in the IAT. All partic- addicted/normal) as a between-subjects factor. Post hoc (LSD)
ipants had normal or corrected-to-normal vision and were free of was used when necessary. Bonferroni correction was used when
neurological or psychiatric disorders. To control the pathological appropriate.
variables that may confuse the results, the psychological test data The reaction time (RT) in Go trials, and the error rate (ER) in
(SCL-90 and 16PF) obtained when the students entered the univer- NoGo trials from both groups were subjected to t-tests. There was
sity were checked. Mental states of the IAD group members prior to no difference between these two groups in Go-RTs (t(1,22) = 0.29,
Internet addiction were verified further by conducting interviews p > 0.05. IAD, 306.2 ± 30.7; normal, 297.9 ± 30.5). There was no
with their classmates and tutors. The experiment procedure is in difference between these two groups in ERs in NoGo condition
accordance with the ethical principles of the 1964 Declaration of (t(1,22) = 0.13, p > 0.05. IAD, 0.088 ± 0.009; normal, 0.090 ± 0.014).
Helsinki (World Medical Organization). No significant difference was found in behavioral performance
Stimuli for the Go/NoGo task were the capital letters “O” and “S.” between these two groups.
Stimuli were presented at the center of the screen. Each trial started Fig. 1 shows the ERP waveforms at Fz, Cz, and Pz electrode sites.
with a small white cross (+) at the center of the screen against a Table 1 shows the mean amplitudes and peak latencies in N2 and
black background for 250 ms, followed by a stimulus letter. Stimuli P3.
were presented for 500 ms and then terminated by a key press. After The significant stimulus main effect in the mean amplitude of N2
a 500 ms black screen appears, the next trial is initiated. During each was larger for NoGo than Go condition in both IAD [F(1,11) = 5.04,
trial, one of the two letters was presented, and either a response p < 0.05] and normal groups [F(1,11) = 5.18, p < 0.05]. Significant
(Go) or the withholding of a response (NoGo) was required. difference was found between IAD and normal groups in NoGo
After an initial practice block of 40 stimuli, four experimental condition, the IAD group elicited significant lower N2 mean ampli-
blocks each consisting of 80 stimuli (25% NoGo probability) were tude than normal group [F(1,22) = 6.92, p < 0.05]. The difference was
completed with 1–2 min breaks between blocks. The “S” and “O” in largest at the central sites (1.21 ␮V), as compared with frontal sites
Go and NoGo cues were counter-balanced among these four blocks. (0.76 ␮V) and parietal sites (0.69 ␮V). No significant difference was
The stimuli procedure and behavioral data were collected using the found between these two groups in Go condition [F(1,22) = 1.21,
E-prime software system. p > 0.05]. The peak latencies in NoGo conditions were significant
Two methods were used to stimulate the impulse of partici- longer than Go conditions in both IAD group [F(1,11) = 5.22, p < 0.05]
pants to respond. First, more Go trials were conducted than NoGo and normal group [F(1,11) = 4.96, p < 0.05]. However, in N2 peak
trials (4:1). Most of the trials were Go cues; thus, when the NoGo latency, group main effect was not found in Go [F(1,22) = 1.35,
cues appeared, participants had to control their impulse to respond, p > 0.05] or NoGo conditions [F(1,22) = 2.19, p > 0.05] when com-
inducing the impulse inhibition process. Second, participants were pared IAD and normal groups. The interaction of group and stimulus
informed before the task that the six best-performing (shorter reac- was not significant in N2 mean amplitude [F(1,22) = 1.41, p > 0.05]
tion time, accurate rate ≥90%) participants would be rewarded with and peak latency [F(1,22) = 1.61, p > 0.05].
an additional 30 dollars. The significant main effect of stimulus indicated P3 amplitudes
Participants were seated in a dimly lit, electrically shielded, were larger for the NoGo than the Go stimulus in IAD [F(1,11) = 7.18,
and sound-attenuated room. Participants were positioned approx- p < 0.05] and normal [F(1,11) = 6.20, p < 0.05] group. Further analysis
imately 100 cm away from a computer screen (DELL 17-in. LCD between groups showed that IAD group showed significant higher
monitor) with horizontal and vertical visual angles of less than P3 amplitude than normal group in NoGo items [F(1,22) = 6.43,
5◦ . All participants were instructed to keep their eyes fixed at the p < 0.05]. No significant main effect was found in Go trials between
center of the screen throughout the entire task. these two groups [F(1,22) = 1.18, p > 0.05]. In peak latencies of P3,
High-density ERPs of each participant were recorded using a IAD group elicited significant longer P3 latency than normal group
128-channel geodesic sensor net (EGI system, Electrical Geodesics in NoGo condition [F(1,22) = 4.88, p < 0.05], but no significant dif-
Inc., Eugene, OR, USA), coupled to a high-input impedance ampli- ference was found in Go condition [F(1,22) = 1.21, p > 0.05]. The
140 G. Dong et al. / Neuroscience Letters 485 (2010) 138–142

Fig. 1. The grand-average ERP waveforms at Fz, Cz, and Pz electrode sites in different groups. N2 and P3 components were found in both groups. The IAD group elicited lower
N2 and higher P3 amplitude than normal group in impulse control condition (NoGo trials).
G. Dong et al. / Neuroscience Letters 485 (2010) 138–142 141

Table 1
Mean amplitude and peak latencies in Go/NoGo conditions in different groups.

Go, mean amplitude NoGo, mean amplitude Go, peak latency NoGo, peak latency

IAD Normal IAD Normal IAD Normal IAD Normal

N2
M 0.87 0.81 1.12 1.43 168.3 177.9 197.6 203.1
SD 0.15 0.15 0.25 0.27 23.1 25.8 33.0 35.4
P3
M 1.18 1.23 1.92 2.42 301.9 304.6 369.8 321.4
SD 0.22 0.24 0.41 0.46 50.1 56.6 57.3 56.5

Notes: M, arithmetic mean; SD, standard deviation; IAD, Internet addiction disorder group; Normal, normal group.

interaction of group and stimulus was not significant in P3 mean endeavors for participants to successfully inhibit their response
amplitude [F(1,22) = 1.87, p > 0.05] and peak latency [F(1,22) = 2.02, impulses.
p > 0.05]. The relation between mean amplitude (P3) and the ER The NoGo-P3 latency was longer in IAD-afflicted participants
in NoGo condition was 0.053 (p > 0.05) in IAD and 0.048 (p > 0.05) compared with that of normal subjects. Peak latency is associated
in normal group. In RTs, the relation between mean amplitude (P3) with cognitive efficiency. P3 latency is an indicator of processing
and RT in Go conditions was 0.076 (p > 0.05) in IAD group and 0.068 speed [24,26] suggesting that IAD-afflicted students had less effi-
(p > 0.05) in normal group. No significant correlation was found cient information processing function than their normal peers. On
during these analyses. the other hand, the longer P3 amplitude may be related to impaired
In the present study, ERPs were used to investigate the response impulse control. Evidence from studies on impaired inhibitory abil-
inhibition difference between IAD-afflicted and normal students ity shows that PTSD and Parkinson’s disease groups have longer
during a Go/NoGo task, by focusing on the NoGo-N2 and NoGo- NoGo-P3 latency compared with control groups [4,27].
P3 components related to the impulse inhibition. In the behavioral In summary, IAD-afflicted participants displayed less efficient
performances, no significant difference was found between the IAD brain function not only with respect to information processing, but
and normal groups. This may indicate that behavioral results were also response inhibition.
not sensitive enough to allow for measurement of the difference Taking all features of N2 and P3 components into considera-
between the two groups. Our main findings could be summarized tion, we can comprehensively understand impulse control in the
as follows. The mean amplitude was lower in N2, but higher in P3 IAD-afflicted students. N2 may represent the detection of response
in NoGo conditions in the IAD group compared with the normal conflict [14] or recognition of the need for inhibition [28]. NoGo-
group. The peak latency was longer in P3 in NoGo conditions in the P3 is associated with response evaluation or successful response
IAD group than in the normal group. inhibition [6]. Peak latency is associated with processing efficiency
All participants (IAD and normal) showed a distinct NoGo-N2 during the cognitive processes [24,26]. The IAD-afflicted students
effect, that is, the NoGo stimuli elicited larger N2 amplitude com- showed a lower activation in detecting response conflict (lower N2
pared with the Go stimuli. The NoGo-N2 effect reported in this amplitude) than normal students. As a result, they had to engage in
study is consistent with that of previous studies [14,16,20]. How- more cognitive endeavors to complete the inhibition task (higher
ever, the NoGo-N2 in the IAD group was significantly lower than P3 amplitude). They were less efficient during these processes than
in the normal group. Findings from previous Go/NoGo studies the normal participants (longer P3 peak latency).
support the hypothesis that the NoGo-N2 is driven by inhibi- Limitations of the present research should be noted. First, all the
tion of a planned response [18]. Researchers also believe that participants were right-handed students as is the case in most ERP
the NoGo-N2 reflects the conflict monitoring process, and the studies. Second, all the participants were male students because
increase of NoGo-N2 amplitude could be interpreted as reflecting few females are addicted to the Internet. Consequently, our findings
higher demands on inhibitory control [22]. The NoGo-N2 amplitude only apply to right-handed male individuals.
could reflect the association between amplitude and successful The present results show that the conflict detection ability of
response inhibition [17]. In summary, the N2 may be used to IAD-afflicted students was impaired as recorded by indexing of
detect conflict and the insistence of participants on controlling their the lower NoGo-N2 mean amplitude. These students had lower
impulses. inhibitory ability and slower speed as indexed by higher NoGo-
In the present study, the IAD group showed lower N2 amplitude P3 amplitude and shorter NoGo-P3 latency. The inferior response
than the normal group, suggesting that their attention and ability inhibition ability of the IAD group might have come from the first
to detect conflict were impaired. The NoGo stimuli elicited signifi- stages of conflict detection to the later stage of response evaluation
cantly higher P3 amplitude than the Go stimuli in all participants. or successful response inhibition. This result provides evidence for
This result can be explained by the characteristics of the present the assumption that inhibitory ability is impaired in IAD-afflicted
research task. The number of Go and NoGo trials was in the 4:1 people.
ratio; thus, the NoGo trials are similar to the infrequent targets in
oddball tasks [12]. Acknowledgement
The IAD group elicited higher P3 amplitude than the nor-
mal group. The posterior P3 is an index of the inhibition of This research was supported by National Science Foundation of
task-irrelevant information, and representation of later conscious China (30900405).
categorization, decision-making, and premotor response-related
activities [7,8,19]. NoGo-P3 is also believed to be an index of References
response inhibition [6,14,28]. Thus, the size of P3 amplitudes in the
present experiment might reflect the degree of cognitive endeav- [1] G.M. Barnes, J.W. Welte, J.H. Hoffman, B.A. Dintcheff, Shared predictors of
ors when the participants successfully inhibited their impulse to youthful gambling, substance use, and delinquency, Psychology of Addictive
Behaviors 19 (2005) 165–174.
respond. The IAD group elicited higher P3 amplitude than the nor- [2] K.W. Beard, E.M. Wolf, Modification in the proposed diagnostic criteria for
mal group. This result may reflect the need for more cognitive Internet addiction, Cyberpsychology and Behavior 4 (2001) 377–383.
142 G. Dong et al. / Neuroscience Letters 485 (2010) 138–142

[3] A.C. Bedard, A. Ickowicz, G.D. Logan, S. Hogg-Johnson, R. Schachar, R. Tannock, [18] J.R. Folstein, C. Van Petten, Influence of cognitive control and mismatch on the
Selective inhibition in children with attention-deficit hyperactivity disorder off N2 component of the ERP: a review, Psychophysiology 45 (2008) 152–170.
and on stimulant medication, Journal of Abnormal Child Psychology 31 (2003) [19] A. Goldstein, K.M. Spencer, E. Donchin, The influence of stimulus deviance
315–327. and novelty on the P300 and Novelty P3, Psychophysiology 39 (2002)
[4] H. Bokura, S. Yamaguchi, S. Kobayashi, Event-related potentials for response 781–790.
inhibition in Parkinson’s disease, Neuropsychologia 43 (2005) 967–975. [20] S.J. Johnstone, C.B. Pleffer, R.J. Barry, A.R. Clarke, J.L. Smith, Development of
[5] K.J. Bruin, A.A. Wijers, Inhibition, response mode, and stimulus probability: inhibitory processing during the Go/NoGo task: a behavioral and event-related
a comparative event-related potential study, Clinical Neurophysiology 113 potential study of children and adults, Journal of Psychophysiology 19 (2005)
(2002) 1172–1182. 11–23.
[6] K.J. Bruin, A.A. Wijers, A.S.J. Van Staveren, Response priming in a go/nogo task: [21] L.M. Jonkman, The development of preparation, conflict monitoring and inhi-
do we have to explain the go/nogo N2 effect in terms of response activation bition from early childhood to young adulthood; a Go/Nogo ERP study, Brain
instead of inhibition? Clinical Neurophysiology 112 (2001) 1660–1671. Research 1097 (2006) 181–193.
[7] S. Campanella, C. Gaspard, D. Debatisse, R. Bruyer, M. Crommelinck, J.M. Guerit, [22] J.L. Kenemans, E.M. Bekker, M. Lijffijt, C.C. Overtoom, L.M. Jonkman, M.N.
Discrimination of emotional facial expressions in a visual oddball task: an ERP Verbaten, Attention deficit and impulsivity: selecting, shifting, and stopping,
study, Biological Psychology 59 (2002) 171–186. International Journal of Psychophysiology 58 (2005) 59–70.
[8] S. Campanella, M. Rossignol, S. Mejias, F. Joassin, P. Maurage, D. Debatisse, R. [23] E. Kirmizi-Alsan, Z. Bayraktaroglu, H. Gurvit, Y.H. Keskin, M. Emre, T. Demi-
Bruyer, M. Crommelinck, J.M. Guérit, Human gender differences in an emotional ralp, Comparative analysis of event-related potentials during Go/NoGo and
visual oddball task: an event-related potentials study, Neuroscience Letters 367 CPT: decomposition of electrophysiological markers of response inhibition and
(2004) 14–18. sustained attention, Brain Research 1104 (2006) 114–128.
[9] F. Cao, L. Su, T. Liu, X. Gao, The relationship between impulsivity and Internet [24] L.K. McEvoy, E. Pellouchoud, M.E. Smith, A. Gevins, Neurophysiological sig-
addiction in a sample of Chinese adolescents, European Psychiatry 22 (2007) nals of working memory in normal aging, Cognitive Brain Research 11 (2001)
466–471. 363–376.
[10] P. Cavedini, G. Riboldi, R. Keller, A. D’Annucci, L. Bellodi, Frontal lobe dysfunction [25] F.G. Moeller, D.M. Dougherty, E.S. Barratt, J.M. Schmitz, A.C. Swann, J.
in pathological gambling patients, Biological Psychiatry 51 (2002) 334–341. Grabowski, The impact of impulsivity on cocaine use and retention in treat-
[11] R.A. Davis, Cognitive-behavioral model of pathological Internet use, Computers ment, Journal of Substance Abuse Treatment 21 (2001) 193–198.
in Human Behavior 17 (2001) 187–195. [26] J. Polich, J.R. Criado, Neuropsychology and neuropharmacology of P3a and P3b,
[12] E. Donchin, Surprise!. . .Surprise? Psychophysiology 18 (1981) 493–513. International Journal of Psychophysiology 60 (2006) 172–185.
[13] G. Dong, L. Yang, Y. Hu, Y. Jiang, Is N2 associated with successful suppression [27] J.L. Shucard, D.C. McCabe, H. Szymanski, An event-related potential study
of behavior responses in impulse control processes? Neuroreport 20 (2009) of attention deficits in posttraumatic stress disorder during auditory and
537–542. visual Go/NoGo continuous performance tasks, Biological Psychology 79 (2008)
[14] F.C.L. Donkers, G.J.M. Van Boxtel, The N2 in go/no-go tasks reflects conflict 223–233.
monitoring not response inhibition, Brain and Cognition 56 (2004) 165–176. [28] J.L. Smith, S.J. Johnstone, R.J. Barry, Movement-related potentials in the
[15] A.C. Douglas, J.E. Mills, M. Niang, S. Stepchenkova, S. Byun, C. Ruffini, S.K. Lee, Go/NoGo task: the P3 reflects both cognitive and motor inhibition, Clinical
J. Loutfi, J.-K. Lee, M. Atallah, M. Blanton, Internet addiction: meta-synthesis of Neurophysiology 119 (2008) 704–714.
qualitative research for the decade 1996–2006, Computers in Human Behavior [29] F. Vitaro, L. Arseneault, R.E. Tremblay, Impulsivity predicts problem gambling
24 (2008) 3027–3044. in low SES adolescent males, Addiction 94 (1999) 565–575.
[16] X. Duan, J. Shi, J. Wu, Y. Mou, H. Cui, G. Wang, Electrophysiological corre- [30] K.S. Young, R.C. Rogers, The relationship between depression and Internet
lates for response inhibition in intellectually gifted children: a Go/NoGo study, addiction, Cyberpsychology and Behavior 1 (1998) 25–28.
Neuroscience Letters 457 (2009) 45–48. [31] B.W. Zhang, L. Zhao, J. Xu, Electrophysiological activity underlying inhibitory
[17] M. Falkenstein, J. Hoormann, J. Hohnsbein, ERP components in Go/Nogo tasks control processes in late-life depression: a Go/Nogo study, Neuroscience Letters
and their relation to inhibition, Acta Psychologica 101 (1999) 267–291. 419 (2007) 225–230.

You might also like