Neuroscience 124 (2004) 953–961
LEADING ROLE OF THE PIRIFORM CORTEX OVER THE NEOCORTEX
IN THE GENERATION OF SPONTANEOUS INTERICTAL SPIKES
DURING BLOCK OF GABAA RECEPTORS
P. RIGAS AND M. A. CASTRO-ALAMANCOS*
Department of Neurobiology and Anatomy, Drexel University College
of Medicine, 2900 Queen Lane, Philadelphia, PA 19129, USA
Abstract—Interictal spikes are generated in the cerebral cor-
tex during certain pathological conditions. In normal tissue,
interictal spikes are triggered by blocking GABAA receptors.
We studied the propensity of different areas of the cerebral
cortex (including neocortex and piriform cortex) to generate
spontaneous interictal spikes during block of GABAA recep-
tors in slices of adult mice. Ten sequential brain slices were
studied spanning most of the cerebral hemisphere. During
block of GABAA receptors spontaneous interictal spikes
were observed in all slices. Interestingly, interictal spikes
recurred at different frequencies in different slices; posterior
slices had a higher rate than the frontal slices (approximately
0.2 vs. 0.1 Hz, posterior vs. frontal). Multi-site recordings
allowed to monitor discharges traveling across each slice
and to derive cross-correlations. It became apparent that for
the slices with higher frequency (i.e. posterior slices) the
discharges originated in the piriform cortex and spread to the
neocortex. A correlation between the frequency of the spon-
taneous discharges and the probability of events originating
from piriform cortex was positive and significant. A further
demonstration of this site of origin was provided by severing
the connections between the neocortex and the piriform cor-
tex. After interrupting these connections all the neocortical
sites in the posterior slices displayed a lower frequency of
discharges; similar to slower frontal slices. Meanwhile, the
isolated piriform cortex of the posterior slices continued to
produce higher frequency discharges. Thus, the higher fre-
quency activity displayed by the posterior slices is intrinsi-
cally generated in the posterior piriform cortex from where it
spreads to the neocortex. The results indicate that the neo-
cortex and the piriform cortex have a distinct propensity to
generate spontaneous interictal spikes, and that the more
prone areas impose their activity on the less prone areas
during disinhibition. © 2004 IBRO. Published by Elsevier Ltd.
All rights reserved.
Key words: epilepsy, paroxysmal depolarizing shift,
disinhibition, epileptiform discharge.
Different brain regions have different propensities to
generate epileptiform activity after the application of
chemoconvulsants, high-frequency electrical stimula-
tion, or other manipulations (for review: De Curtis and
Avanzini, 2001; Prince and Connors, 1986; Traub et al.,
1996; Jefferys, 1990; McNamara, 1994). The basic form
*Corresponding author. Fax: ⫹1-215-843-9082.
E-mail address: manuel.castro@drexel.edu (M. Castro-Alamancos).
Abbreviations: ACSF, artificial cerebrospinal fluid; ANOVA, analysis of
variance; BMI, bicuculline.
0306-4522/04$30.00⫹0.00 © 2004 IBRO. Published by Elsevier Ltd. All rights reserved.
doi:10.1016/j.neuroscience.2003.11.034
953
of epileptiform activity is the interictal spike, which in the
depth of the cortex appears as a negative spike followed
by a positive wave. Spike-wave discharges or interictal
spikes (these terms are used interchangeably here) cor-
respond intracellularly to an overt depolarization termed
a paroxysmal depolarizing shift (Matsumoto and Aj-
mone-Marsan, 1964b; Schwartzkroin and Prince, 1978;
Gutnick et al., 1982). Particular interest has been de-
voted to investigating epileptiform events induced by
chemoconvulsants (e.g. GABAA receptor antagonists) in
different parts of the cerebral cortex, such as the neo-
cortex, the piriform cortex and the hippocampus. In the
hippocampus, epileptiform events have been shown to
originate in the CA3 region from where they spread to
CA1 (Schwartzkroin and Prince, 1978; Traub et al.,
1993a,b, 1996; Miles et al., 1984; Hablitz, 1984). The
piriform cortex appears to have an unusually strong
tendency to generate epileptiform events (Hoffman and
Haberly, 1991, 1993, 1996; McIntyre and Wong, 1986;
Piredda and Gale, 1985; Piredda et al., 1985; Stevens et
al., 1988). In the neocortex, interictal spikes are also
generated by blocking GABAA receptors (Connors et al.,
2001; Chagnac-Amitai and Connors, 1989b; Gutnick et
al., 1982; Hablitz, 1987; Ralston, 1958; Matsumoto and
Ajmone-Marsan, 1964a,b; Gloor et al., 1977). In a pre-
vious study, we found that spontaneous rhythmically
recurring spike-wave discharges are induced in slices of
neocortex when GABAA receptors are blocked (Castro-
Alamancos and Rigas, 2002). We also found that after-
discharges that are generated by blocking GABAA and
GABAB receptors preferentially occur in specific cytoar-
chitectonic areas of neocortex, but not in others (Castro-
Alamancos and Rigas, 2002). This raised the possibility
that spike-wave discharges are preferentially generated
in specific cortical areas. In the present study, we ex-
amined the propensity of different regions of the cere-
bral cortex (including neocortex and piriform cortex) to
generate spontaneous interictal spikes during block of
GABAA receptors. Cortical slices were obtained sequen-
tially covering most of the extent of the cerebral hemi-
sphere. Our initial observations revealed that posterior
cortical slices produced spontaneous interictal spikes
with a higher frequency than frontal slices. Cross-corre-
lation analysis and lesion studies demonstrated that the
higher frequency activity of the posterior slices origi-
nated from the piriform cortex. Thus, the posterior piri-
form cortex has a stronger intrinsic propensity to gener-
ate spontaneous interictal spikes than the neocortex.
954 P. Rigas and M. A. Castro-Alamancos / Neuroscience 124 (2004) 953–961
Fig. 1. Photographs of the 10 sequential brain slices used in the present study. Slices are labeled 1–10 from posterior to frontal. The sites that were
recorded are labeled 1–5 from dorsomedial to ventrolateral. Also shown is the location of the cut that interrupted the connections between site 5 and
the other sites (1– 4). Sites 1– 4 correspond to the neocortex, while site 5 corresponds to the piriform cortex.
EXPERIMENTAL PROCEDURES
Slices were prepared from adult (ⱖ7 weeks) BALB/C mice as pre-
viously described (Castro-Alamancos and Rigas, 2002). Mice were
deeply anesthetized with sodium pentobarbitone (60 mg/kg) and
upon losing all responsiveness to a strong tail pinch the brain was
rapidly extracted and placed in ice cold buffer solution. Slices
(400 ␮m thick) were cut in the thalamocortical plane (Agmon and
Connors, 1991) using a vibratome. Six sequential slices of the right
hemisphere were taken from each animal and were kept in an
interface chamber at 32.5 °C. The slices were bathed constantly
(1–1.5 ml/min) with artificial cerebrospinal fluid (ACSF) containing (in
mM): NaCl (126), KCl (3.5), NaH2PO4 (1.25), NaHCO3 (26),
MgSO4䡠7H2O (1), dextrose (10), CaCl2䡠2H2O (1). The ACSF was
bubbled with 95% O2 and 5% CO2. This ACSF composition has been
shown to produce spontaneous slow activity in neocortical slices
(Sanchez-Vives and McCormick, 2000; Castro-Alamancos and Ri-
gas, 2002). All procedures were reviewed and approved by the
Animal Care Committee of McGill University and Drexel University
College of Medicine. The number of animals used in this study was
kept to a minimum and every effort was made to minimize their
suffering.
Field recordings were made using low-impedance pipettes
(approximately 0.5 M⍀) filled with ACSF. Data were stored and
analyzed using Experimenter’s Workbench (Data Wave Tech-
nologies, Longmont, CO, USA) and Origin (Microcal Software,
Northampton, MA, USA) software. Every slice studied was
photographed in the interface chamber using a CCD camera in
order to allow subsequent identification (see Fig. 1).
Statistical analyses were performed using Origin (Microcal Soft-
ware) and Datasim (Bates College, Lewiston, ME, USA) and con-
sisted of t-tests and repeated measures analysis of variance
(ANOVA) with three factors (site/slice/cut) and a Fischer’s test for
post hoc analysis.
RESULTS
Experimental setup and initial observations
Ten sequential slices of the mouse brain spanning most of
the fronto-posterior axis of the right cerebral hemisphere
were studied. Slices were named 1 through 10 from pos-
terior to frontal (Fig. 1). Slice 5 was defined as the first slice
that included an intact thalamocortical connection (i.e. in-
tact thalamocortical axons course from thalamus to neo-
cortex), and the other slices were sequentially labeled
based on that criterion. Moreover, since all slices were
photographed, subsequent identification was possible af-
ter the experiments. Note that since slices are not cut in the
coronal plane they cannot be compared directly with ex-
isting mouse atlases throughout the extent of the slice
(Paxinos and Franklin, 2001). This is because due to the
angled cut in the thalamocortical slices different portions of
the slices will be at different distances from bregma. A
comparison of only the ventrolateral quadrant of the slices
(i.e. area of the piriform cortex) with the atlas of Paxinos
and Franklin (2001) revealed that slice 1 corresponds
roughly to 2.8 mm posterior to bregma and slice 10 corre-
sponds to 0.8 mm anterior to bregma. Due to space and
time constraints six slices were studied per experiment
(animal) and several series of experiments were per-
formed. One series of experiments included the first six
slices (i.e. 1– 6), and another series of experiments in-
cluded the last six slices (i.e. 5–10). Since these two series
of experiments had two common slices (i.e. slices 5 and 6),
they served as a means of controlling for variability be-
tween experiments (see below). Moreover, another series
of experiments included slices 1–3 and 8 –10, which al-
lowed comparing the most frontal and posterior slices in
the same experiment (see below). Recordings were per-
formed from multiple sites (sites 1–5; see Fig. 1) of each
slice (slices 1–10) under control conditions and application
of bicuculline (BMI) or both. The recording sites extended
from the neocortex (sites 1– 4) to the piriform cortex (site
5). Site 4 was generally located around the transition area
between insular cortex and neocortex for the frontal slices
 P. Rigas and M. A. Castro-Alamancos / Neuroscience 124 (2004) 953–961 955

Table 1. Frequency (Hz; mean⫾S.D.) and probability of occurrence (% of slices that displayed spontaneous activity) of spontaneous activity in the
five sites of the 10 slices during control buffer (n⫽number of experiments)

Slice n Site 1 Site 2 Site 3 Site 4 Site 5

1 8 0.04⫾0.03 0.06⫾0.03 0.14⫾0.13 0 0.02⫾0.00

75% 38% 38% 0% 14%
2 9 0.05⫾0.02 0.18⫾0.11 0.17⫾0.12 0.05⫾0.03 0
89% 38% 67% 38% 0%
3 9 0.08⫾0.03 0.07⫾0.06 0.06⫾0.03 0.04⫾0.02 0
56% 44% 44% 22% 0%
4 7 0.02⫾0 0.06⫾0.08 0.03⫾0.02 0.02⫾0.01 0
14% 43% 100% 43% 0%
5 23 0.06⫾0.03 0.11⫾0.08 0.08⫾0.04 0 0.01⫾0.00
35% 30% 18% 0% 5%
6 20 0.12⫾0.07 0.18⫾0.11 0.12⫾0.09 0.01⫾0.00 0.02⫾0.01
60% 59% 55% 7% 15%
7 10 0.10⫾0.08 0.12⫾0.14 0.13⫾0.10 0.03⫾0.00 0.02⫾0.01
70% 100% 80% 17% 22%
8 10 0.11⫾0.08 0.13⫾0.08 0.14⫾0.09 0 0
80% 60% 70% 0% 0%
9 9 0.14⫾0.06 0.10⫾0.07 0.08⫾0.05 0 0.01⫾0.00
33% 67% 78% 0% 11%
10 9 0 0.05⫾0.05 0.12⫾0.06 0.01⫾0.00 0
0% 56% 56% 11% 0%

and between perirhinal cortex and neocortex for the pos-
terior slices. The experimental procedure involved using
two recording electrodes. One electrode was always lo-
cated in site 1, the steady site, while the other electrode
was moved from sites 2–5 recording for 4 –5 min at each
site. A necessary condition for a successful experiment
was that the activity in site 1 was not significantly different
throughout the mapping procedure for each slice (i.e. the
frequency of events in site 1 varied less than 10%).
In control buffer slices displayed spontaneous slow
oscillations, as previously described (Sanchez-Vives and
McCormick, 2000; Castro-Alamancos and Rigas, 2002).
For example, in site 1 these events occurred at a fre-
quency of 0.08⫾0.06 Hz (mean⫾S.D.; n⫽45 slices) and
their amplitude was 0.30⫾0.15 mV. The probability of ob-
serving spontaneous slow oscillations in the neocortex
varied widely between different slices and cortical sites
(see Table 1). On average approximately 51% of the slices
displayed slow oscillations in site 1. There were some
notable exceptions. Site 1 of slice 10 displayed no activity
in control buffer. Also, spontaneous slow oscillations were
found to be routinely absent from site 5 (piriform cortex)
and also quite rare in site 4 of all slices, while they are
usually present in sites 1–3 of all slices. For instance, the
probability of observing spontaneous oscillations in control
buffer was 5% in site 5 of all slices, while the probability in
sites 1–3 of all slices was on average 55%.
Application of a GABAA receptor antagonist (BMI; 10 ␮M)
transformed the slow-wave oscillations into large amplitude
1.7⫾0.4 mV (n⫽93 slices) interictal spikes that recurred
spontaneously, as previously described for frontal slices
(Castro-Alamancos and Rigas, 2002) and in vivo (Castro-
Alamancos, 2000). During BMI, interictal spikes were ob-
served in all slices and cortical sites irrespective of the pres-
ence of spontaneous slow oscillations during control buffer.
Fig. 2 illustrates an example of this activity. A comparison of
the activity recorded from site one of the different slices
revealed that interictal spikes recurred at different frequen-
cies (Fig. 2; see also Table 2). For instance while the frontal
slices 7–10 displayed interictal spikes at a frequency of
0.10⫾0.02 Hz, the posterior slices 1– 6 displayed a signifi-
cantly higher frequency of 0.2⫾0.07 Hz (t-test; P⬍0.0001).
This difference cannot be attributed to the fact that in some
cases the frontal and posterior slices were taken from differ-
ent experiments. Thus, slices 5– 6 were common for two
series of experiments that included either frontal or posterior
slices and the interictal spike frequency was not significantly
different between these two groups of experiments (n⫽20
experiments per group; P⫽0.36). Moreover, in other cases
(n⫽8 animals) we recorded from three frontal slices (8 –10)
and three posterior slices (1–3) within the same experiment
and found that the posterior slices had a significantly higher
frequency than the frontal slices (0.19⫾0.05 vs. 0.11⫾0.2 Hz;
P⬍0.0001). Thus, the posterior slices displayed about double
the interictal spike frequency than the frontal slices (approx-
imately 0.2 vs. 0.1 Hz, posterior vs. frontal). In conclusion,
during block of GABAA receptors posterior slices (slices 1– 6)
produce a significantly higher frequency spontaneous inter-
ictal spike activity than frontal slices (slices 7–10).
Spatiotemporal properties of interictal spikes
During application of BMI inter-ictal spikes occurred in all
cortical sites and slices examined including sites 4 and 5
that were generally silent during control buffer, and also in
site 1 of slice 10, which was also silent (see Table 2). Thus,
BMI produced synchronous activity over all five sites re-
corded, and it seemed that interictal spikes spread from
one site to another. This became apparent using cross-
correlation analysis, which was computed by using thresh-
956 P. Rigas and M. A. Castro-Alamancos / Neuroscience 124 (2004) 953–961

ent sites depending on the slice. For instance, in some

slices activity would consistently start in site 5 and spread
up to site 1 (Fig. 3, upper panel); in other cases activity
would originate in any of the other sites and spread from
there. Fig. 3 shows several examples from different slices
where the activity originated in different sites. It became
apparent that the slices in which site 5 steadily preceded
all the other sites (i.e. more than 80% of the events origi-
nated at site 5) had a higher frequency of spontaneous
interictal spike activity. This is illustrated in Fig. 4. A linear
correlation analysis between the frequency of the sponta-
neous interictal spike activity and the probability of events
originating from site five was positive and significant
(r⫽0.73, P⬍0.0001; n⫽180; Fig. 4, upper panel). Fig. 4
(middle and lower panels) also shows that this relationship
was only present for the posterior slices (slices 1– 6;
r⫽0.63; P⬍0.0001) and not for the frontal slices, which had
a lower frequency of spontaneous interictal spike activity.
Thus, during BMI the higher frequency activity of posterior
slices may originate in the piriform cortex (site 5) from
where it spreads to the neocortex (sites 1– 4).

Effect of interrupting the connection between

Fig. 2. Spontaneous activity in neocortex during application of control
buffer and during BMI. A: Examples of field potential recordings during
control buffer (control) and during the application of a GABAA receptor
antagonist (BMI; 10 ␮M). The upper traces show close-ups of typical
events recorded under both conditions. Recordings were taken from
site 2 of slice 6 but are representative of other neocortical sites. B:
Frequency of spontaneous activity (interictal spikes) recorded from
site 1 of each of the 10 slices studied during application of BMI. Note
the higher frequency of spontaneous interictal spikes in the posterior
slices.
old detectors to establish the timing relationship between
events in each site and those in site 1. Propagation speed
was calculated at around 0.03– 0.07 m/s for every case,
which is similar to previous studies that have shown that
discharges spread through the neocortex and piriform cor-
tex at about 0.03– 0.1 m/s (Connors et al., 2001; Telfeian
and Connors, 1998; Chagnac-Amitai and Connors, 1989a;
Chervin et al., 1988; Demir et al., 1998). Cross-correlations
revealed that the interictal spikes could originate in differ-
piriform cortex and neocortex
Since the higher frequency activity observed in the poste-
rior slices seems to originate in the piriform cortex (site 5)
we reasoned that interrupting the connection between the
piriform cortex and the neocortex would affect the fre-
quency of activity observed in the neocortex. The next
experiment was performed to test this hypothesis by com-
paring the spontaneous activity during BMI in all sites and
slices before and after a cut was performed to interrupt the
connections between site 5 and the other sites 1– 4 (see
Fig. 1). The cut extended from the pia through the white
matter interrupting all connections between both cortical
regions. A three-way ANOVA of the spontaneous interictal
spike frequency was performed. The three factors used
were: the “slice,” the “site” and the effect of the “cut.” Fig.
5 shows that in intact slices the frequency of spontaneous
interictal spike activity for all sites within a slice was fairly
regular because the activity would spread between sites.
Thus, “site” was not a significant factor in the ANOVA
(P⫽0.91). The frontal slices (7–10) had a significantly
lower frequency of spontaneous interictal spike activity

Table 2. Frequency (Hz; mean⫾S.D.) of spontaneous interictal spikes in the five sites of the 10 slices during block of GABAA receptors with BMI
(n⫽number of experiments)

Slice n Site 1 Site 2 Site 3 Site 4 Site 5

1 30 0.17⫾0.07 0.17⫾0.06 0.17⫾0.07 0.17⫾0.07 0.17⫾0.07

2 26 0.17⫾0.06 0.17⫾0.06 0.17⫾0.06 0.17⫾0.06 0.17⫾0.06
3 24 0.19⫾0.07 0.18⫾0.07 0.18⫾0.07 0.19⫾0.08 0.19⫾0.07
4 23 0.18⫾0.08 0.17⫾0.07 0.18⫾0.08 0.20⫾0.13 0.20⫾0.13
5 36 0.21⫾0.05 0.19⫾0.05 0.21⫾0.06 0.24⫾0.08 0.25⫾0.08
6 33 0.13⫾0.05 0.13⫾0.04 0.13⫾0.05 0.13⫾0.05 0.14⫾0.05
7 29 0.11⫾0.02 0.11⫾0.02 0.11⫾0.02 0.11⫾0.02 0.11⫾0.02
8 27 0.11⫾0.02 0.11⫾0.02 0.11⫾0.02 0.11⫾0.02 0.11⫾0.02
9 22 0.10⫾0.02 0.10⫾0.02 0.10⫾0.02 0.10⫾0.02 0.10⫾0.02
10 19 0.11⫾0.01 0.11⫾0.01 0.11⫾0.01 0.11⫾0.01 0.11⫾0.01
 P. Rigas and M. A. Castro-Alamancos / Neuroscience 124 (2004) 953–961 957

Fig. 3. Examples of cross-correlations from different slices in which the spontaneous interictal spikes originated from different sites. The timing of
events at each site are plotted with respect to site 1 (time zero) and each site is color coded. Upper, In this slice the events would consistently (100%
of events) start around site 5 (red) and spread from there to site 1 (black at 0) by passing through sites 4, 3 and 2. This example corresponds to a
slice 5. Middle, In this slice the events would consistently start around site 1 (100% of events) and spread to site 5. This example corresponds to a
slice 9. Lower, In this slice the events would either start around site 5 (35% of events) and spread to site 1 or events would start around site 1 (65%
of events) and spread from there to site 5. This example corresponds to a slice 4.

than the posterior slices (2–5), and this was reflected in the
significant effect of the factor “slice” (P⬍0.0001). More-
over, this was significant for all five sites examined
(P⬍0.005). It was also apparent that in some of the pos-
terior slices (i.e. slices 4 –5) the frequency was higher in
the piriform cortex area (sites 4 –5), suggesting that these
sites might be the origin of the higher frequency activity in
those slices; this was reflected in a significant two-way
interaction between “slice” and “site” (P⬍0.0001). After
interrupting the cortical connection between site 5 and the
other neocortical sites by performing a cut between sites 4
and 5, all the posterior slices (1– 6) displayed a lower
958 P. Rigas and M. A. Castro-Alamancos / Neuroscience 124 (2004) 953–961
Fig. 4. Linear correlation analysis between the frequencies of spon-
taneous interictal spike activity and the probability of events originating
from site 5 determined using cross-correlations analysis (see Fig. 3).
Correlation analysis for all the slices included in the study (upper), and
for posterior slices 1– 6 only (middle) and frontal slices 7–10 only
(lower). A regression line is used to fit each group of data.
frequency of spontaneous interictal spike activity in sites
1– 4, while the higher frequency activity in site 5 was
unaffected (Fig. 5). In fact, after the cut sites 1– 4 of the
posterior slices displayed a frequency of spontaneous in-
terictal spike activity that is quite similar to the activity of
frontal slices. Statistically, this was reflected in a significant
main effect of the factor “cut” (P⬍0.0001) and a significant
two-way interaction between “slice” and “cut” (P⬍0.0001).
These results indicate that the higher frequency activity
displayed by the posterior slices is intrinsically generated
in the piriform cortex from where it spreads to the
neocortex.
The ANOVA also showed a three-way interaction
(slices⫻sites⫻cut). The cut differently affected the activity
of sites 1–5 in posterior and anterior slices. While the cut
significantly decreased the activity of sites 1– 4 in posterior
slices 1–5 (P⬍0.05), it had no effect on the activity of the
same sites in slices 7–10. Thus, after the cut the neocor-
tical activity of the posterior slices resembles the low ac-
tivity of the frontal slices (approximately 0.1 Hz). Moreover,
the degree of significance at which the activity of sites 1– 4
in the posterior slices was reduced differed; slices 3–5
were more strongly affected (P⬍0.001) than slice 2
(Pⱕ0.01) and slice 1 which was only marginally affected
(P⫽0.04). On the contrary, in the frontal slices (7–10) the
Fig. 5. Effect of interrupting the connection between neocortex (site 4)
and piriform cortex (site 5) on the spontaneous interictal spikes gen-
erated during BMI in each of the 10 slices and 5 sites studied. A: The
plots show the frequency of the spontaneous interictal spikes during
BMI for every recording site and slice in intact slices (upper) and in
those same slices after a cut interrupting the connection between site
5 and site 4 (lower). The data are mean⫾S.D. from 88 slices (n1⫽15;
n2⫽12; n3⫽11; n4⫽10; n5⫽10; n6⫽9; n7⫽6; n8⫽5; n9⫽6; n10⫽4).
B: Examples of recordings from several sites and slices before and
after a cut interrupting the connections between site 5 and site 4.
Upper, Site 2 and site 5 are recorded simultaneously in a posterior
slice (slice 5; upper left) before (intact) and after the cut (cut; upper
right). Note that during intact conditions the activity is of high frequency
(approximately 0.2 Hz) in both sites 2 (neocortex) and 5 (piriform
cortex). However, after the cut the activity in site 2 becomes signifi-
cantly reduced (approximately 0.1 Hz) while the activity in site 5 is of
similar frequency as before the cut (approximately 0.2 Hz). Lower, The
same is shown for a frontal slice (slice 9). Before the cut (intact) the
activity is of low frequency in both sites 2 and 5 of slice 9 (lower left),
while after the cut the activity in site 2 is not significantly changed while
the activity in site 5 is abolished.
activity in sites 1– 4 did not change after the cut, while the
activity in site 5 was abolished after the cut, indicating that
the spontaneous interictal spikes observed in that site
were propagated from sites 1– 4.
Cross-correlation analysis during BMI revealed that
before the cut activity in most of the posterior slices tended
to start in site 5 (i.e. piriform cortex) as described above
 P. Rigas and M. A. Castro-Alamancos / Neuroscience 124 (2004) 953–961 959

Fig. 6. Origin of spontaneous interictal spike activity in intact slices (A) and after a cut interrupting the connections between sites 4 and 5 (B). Plotted
is the number of slices, as a percentage of the total (%), in which the majority (⬎80%) of all the events began at the site indicated in the x axis. If none
of the sites met this criterion then the slice would be included in the column X, indicating that none of the sites were the preferred origin. Each site
is represented by a distinct column. After the cut (B) site 5 was eliminated from the plot because events could not spread from there to other sites.

(Fig. 6). Slices were classified according to the site of
origin by counting the number of events that started at
each site. For example, a slice was classified as a site 1
slice if more than 80% of the events started at site 1. If
none of the sites met this criterion then the slice was
classified as X. Fig. 6 shows that in more than 75% of
slices 2–5 spontaneous events during BMI originated in
site 5 (Fig. 6A). Thus, the events in the posterior slices
tended to originate in site 5. In contrast, the frontal slices
tended not to have a particular site of origin (i.e. most were
classified as X), or site 2 was generally the site of origin of
the events. After the cut severing the connections between
sites 4 and 5 activity in the posterior slices (2– 6) did not
generally have a preferred site of origin and most were
classified in group X. These results demonstrate that the
higher frequency spontaneous interictal spike activity ob-
served throughout all sites of the posterior slices during
BMI originates in the piriform cortex (site 5).
DISCUSSION
The present study found that application of a GABAA re-
ceptor antagonist produces spontaneous interictal spikes
that recur at different frequencies in different slices. Thus,
960 P. Rigas and M. A. Castro-Alamancos / Neuroscience 124 (2004) 953–961
frontal slices (termed here slices 7–10) produced signifi-
cantly lower frequency discharges than posterior slices
(slices 1– 6). Cross-correlation analysis of the propagating
discharges within each slice suggested that the higher
frequency activity of posterior slices originated from site 5
in those slices, which corresponds to the piriform cortex. A
lesion that interrupted the connections between the piri-
form cortex and the neocortex (i.e. between sites 4 and 5)
demonstrated that the higher frequency activity is intrinsi-
cally generated in the piriform cortex from where it drives
activity in the neocortex. In conclusion, both the neocortex
and the piriform cortex have the intrinsic ability to generate
interictal spikes during block of GABAA receptors, but be-
cause of the intrinsic capability of piriform cortex to gener-
ate higher frequency discharges it drives the activity in the
neocortex, which has a propensity to generate discharges
at a lower frequency. Thus, during certain pathological
conditions the piriform cortex may be a principal site for the
origination of interictal spikes that can drive such activity in
the neocortex.
The present study shows that both neocortex and piri-
form cortex have the intrinsic capability to generate spon-
taneous interictal spikes during block of GABAA receptors.
However, the piriform cortex has a propensity to generate
such activity at a higher frequency, about double the ac-
tivity of the fastest neocortical areas. Epileptiform activity in
the piriform cortex has been shown to arise from the en-
dopiriform nucleus (Hoffman and Haberly, 1991, 1993,
1996; Piredda and Gale, 1985; Piredda et al., 1985;
Stevens et al., 1988). Thus, it is reasonable to assume that
in the present study the faster activity of the posterior slices
originated within the endopiriform nucleus. The present
findings also revealed that there is a difference between
different parts of the piriform cortex. For instance, the
frontal part of the piriform cortex did not generate faster
interictal spikes than the neocortex, and when it was dis-
connected from the neocortex it did not generate sponta-
neous interictal spikes at all. It was only the posterior slices
that produced a faster interictal spike generation and this
activity persisted after disconnection from the neocortex. In
fact, the piriform cortex has been divided into anterior and
posterior parts based on cytoarchitectonic criteria (Haberly
and Price, 1978a,b). The borderline between anterior and
posterior piriform cortex is defined by the disappearance of
the lateral olfactory tract, a concomitant increase in layer III
and the appearance of the ventral endopiriform nucleus.
This division roughly matches the separation between an-
terior and posterior slices in the present study. Interest-
ingly, kindling studies in vivo have suggested that the
posterior piriform cortex, and in particular the deep cell
layer of the rostral portion of the posterior piriform cortex,
plays a critical role in kindling induced seizures (Honack et
al., 1991; Wahnschaffe et al., 1993). However, administra-
tion of BMI in different parts of the piriform cortex revealed
no differences in seizure susceptibility, or a tendency of the
anterior piriform cortex to show higher susceptibility in
some species (Piredda and Gale, 1985; Piredda et al.,
1985; Ebert et al., 2000). The extensive associative fiber
system in the piriform cortex may explain the discrepancy
between electrically and chemically induced seizures. It is
also interesting that the transition area between anterior
and posterior piriform cortex is the piriform area with both
higher density and total number of GABAergic interneu-
rons (Loscher et al., 1998). These characteristics of the
posterior piriform cortex may provide an explanation for the
higher susceptibility of this region to generate interictal
spikes after BMI.
Another interesting finding in the present study was
that while the piriform cortex presented the highest fre-
quency of spontaneous activity during blockade of GABAA
receptors, it generally did not generate spontaneous activ-
ity in control buffer, while the neocortical sites did. We
speculate that the endopiriform nucleus and the overlying
upper layers (I–III) of the piriform cortex are incapable of
generating slow oscillations, which may be an exclusive
capability of the neocortex because of the presence of the
infragranular layers (layers V and VI).
In conclusion, the present study shows the stronger
propensity of the posterior piriform cortex to generate
spontaneous interictal spikes during block of GABAA re-
ceptors, as compared with any neocortical area.
REFERENCES
Agmon A, Connors BW (1991) Thalamocortical responses of mouse
somatosensory (barrel) cortex in vitro. Neuroscience 41:365–379.
Castro-Alamancos MA (2000) Origin of synchronized oscillations in-
duced by neocortical disinhibition in vivo. J Neurosci 20:9195–
9206.
Castro-Alamancos MA, Rigas P (2002) Synchronized oscillations
caused by disinhibition in rodent neocortex are generated by recur-
rent synaptic activity mediated by AMPA receptors. J Physiol 542:
567–581.
Chagnac-Amitai Y, Connors BW (1989a) Horizontal spread of syn-
chronized activity in neocortex and its control by GABA-mediated
inhibition. J Neurophysiol 61:747–758.
Chagnac-Amitai Y, Connors BW (1989b) Synchronized excitation and
inhibition driven by intrinsically bursting neurons in neocortex.
J Neurophysiol 62:1149 –1162.
Chervin RD, Pierce PA, Connors BW (1988) Periodicity and direction-
ality in the propagation of epileptiform discharges across neocortex.
J Neurophysiol 60:1695–1713.
Connors BW, Pinto DJ, Telfeian AE (2001) Local pathways of seizure
propagation in neocortex. Int Rev Neurobiol 45:527–546.
De Curtis M, Avanzini G (2001) Interictal spikes in focal epileptogen-
esis. Prog Neurobiol 63:541–567.
Demir R, Haberly LB, Jackson MB (1998) Voltage imaging of epilep-
tiform activity in slices from rat piriform cortex: onset and propaga-
tion. J Neurophysiol 80:2727–2742.
Ebert U, Wlaz P, Loscher W (2000) High susceptibility of the anterior
and posterior piriform cortex to induction of convulsions by bicucul-
line. Eur J Neurosci 12:4195–4205.
Gloor P, Quesney LF, Zumstein H (1977) Pathophysiology of gener-
alized penicillin epilepsy in the cat: the role of cortical and subcor-
tical structures: II. Topical application of penicillin to the cerebral
cortex and to subcortical structures. Electroencephalogr Clin Neu-
rophysiol 43:79 –94.
Gutnick MJ, Connors BW, Prince DA (1982) Mechanisms of neocor-
tical epileptogenesis in vitro. J Neurophysiol 48:1321–1335.
Haberly LB, Price JL (1978a) Association and commissural fiber sys-
tems of the olfactory cortex of the rat. J Comp Neurol 178:711–740.
Haberly LB, Price JL (1978b) Association and commissural fiber sys-
tems of the olfactory cortex of the rat: II. Systems originating in the
olfactory peduncle. J Comp Neurol 181:781–807.
 P. Rigas and M. A. Castro-Alamancos / Neuroscience 124 (2004) 953–961
Hablitz JJ (1984) Picrotoxin-induced epileptiform activity in
hippocampus: role of endogenous versus synaptic factors. J Neu-
rophysiol 51:1011–1027.
Hablitz JJ (1987) Spontaneous ictal-like discharges and sustained
potential shifts in the developing rat neocortex. J Neurophysiol
58:1052–1065.
Hoffman WH, Haberly LB (1991) Bursting-induced epileptiform EPSPs
in slices of piriform cortex are generated by deep cells. J Neurosci
11:2021–2031.
Hoffman WH, Haberly LB (1993) Role of synaptic excitation in the
generation of bursting-induced epileptiform potentials in the en-
dopiriform nucleus and piriform cortex. J Neurophysiol 70:2550 –
2561.
Hoffman WH, Haberly LB (1996) Kindling-induced epileptiform poten-
tials in piriform cortex slices originate in the underlying endopiriform
nucleus. J Neurophysiol 76:1430 –1438.
Honack D, Wahnschaffe U, Loscher W (1991) Kindling from stimula-
tion of a highly sensitive locus in the posterior part of the piriform
cortex: comparison with amygdala kindling and effects of antiepi-
leptic drugs. Brain Res 538:196 –202.
Jefferys JG (1990) Basic mechanisms of focal epilepsies. Exp Physiol
75:127–162.
Loscher W, Lehmann H, Ebert U (1998) Differences in the distribution
of GABA- and GAD-immunoreactive neurons in the anterior and
posterior piriform cortex of rats. Brain Res 800:21–31.
Matsumoto H, Ajmone-Marsan C (1964a) Cortical cellular phenomena
in experimental epilepsy: ictal manifestations. Exp Neurol 9:305–
326.
Matsumoto H, Ajmone-Marsan C (1964b) Cortical cellular phenomena
in experimental epilepsy: interictal manifestations. Exp Neurol
9:286 –304.
McIntyre DC, Wong RK (1986) Cellular and synaptic properties of
amygdala-kindled pyriform cortex in vitro. J Neurophysiol 55:1295–
1307.
McNamara JO (1994) Cellular and molecular basis of epilepsy. J Neu-
rosci 14:3413–3425.
Miles R, Wong RK, Traub RD (1984) Synchronized afterdischarges in
(Accepted 21 November 2003)
961
the hippocampus: contribution of local synaptic interactions. Neu-
roscience 12:1179 –1189.
Paxinos G, Franklin KBJ (2001) The mouse brain in stereotaxic coor-
dinates. San Diego: Elsevier.
Piredda S, Gale K (1985) A crucial epileptogenic site in the deep
prepiriform cortex. Nature 317:623–625.
Piredda S, Lim CR, Gale K (1985) Intracerebral site of convulsant
action of bicuculline. Life Sci 36:1295–1298.
Prince DA, Connors BW (1986) Mechanisms of interictal epileptogen-
esis. Adv Neurol 44:275–299.
Ralston BL (1958) The mechanism of transition of interictal spiking foci
into ictal seizure discharges. Electroencephalogr Clin Neurophysiol
10:217–232.
Sanchez-Vives MV, McCormick DA (2000) Cellular and network
mechanisms of rhythmic recurrent activity in neocortex. Nat Neu-
rosci 3:1027–1034.
Schwartzkroin PA, Prince DA (1978) Cellular and field potential prop-
erties of epileptogenic hippocampal slices. Brain Res 147:117–130.
Stevens JR, Phillips I, de Beaurepaire R (1988) gamma-Vinyl GABA in
endopiriform area suppresses kindled amygdala seizures. Epilep-
sia 29:404 –411.
Telfeian AE, Connors BW (1998) Layer-specific pathways for the
horizontal propagation of epileptiform discharges in neocortex. Epi-
lepsia 39:700 –708.
Traub RD, Borck C, Colling SB, Jefferys JG (1996) On the structure of
ictal events in vitro. Epilepsia 37:879 –891.
Traub RD, Jefferys JG, Miles R (1993a) Analysis of the propagation of
disinhibition-induced after-discharges along the guinea-pig hip-
pocampal slice in vitro. J Physiol (Lond) 472:267–287.
Traub RD, Miles R, Jefferys JG (1993b) Synaptic and intrinsic con-
ductances shape picrotoxin-induced synchronized after-dis-
charges in the guinea-pig hippocampal slice. J Physiol (Lond)
461:525–547.
Wahnschaffe U, Ebert U, Loscher W (1993) The effects of lesions of
the posterior piriform cortex on amygdala kindling in the rat. Brain
Res 615:295–303.