Sleep Medicine 6 (2005) 2936

www.elsevier.com/locate/sleep

Original article
Topographic mapping of the spectral components
of the cyclic alternating pattern (CAP)
Raffaele Ferria,*, Oliviero Brunib, Silvia Mianoa, Mario G. Terzanoc
a
Department of Neurology I.C., Sleep Research Centre, Oasi Institute for Research on Mental Retardation and Brain Aging (IRCCS), Troina, Italy
b
Department of Developmental Neurology and Psychiatry, Centre for Pediatric Sleep Disorders, University of Rome La Sapienza, Rome, Italy
c
Department of Neurology, Sleep Disorders Center, University of Parma, Parma, Italy
Received 30 March 2004; received in revised form 28 June 2004; accepted 30 June 2004

Abstract

Background and purpose: The aim of this study was to define quantitatively the spectrum content of the sleep pattern termed cyclic
alternating pattern (CAP) A phases, their scalp topography and their probable cortical generators, by using data from sleep polygraphic
recordings that included a large number of scalp EEG channels.
Patients and methods: Polysomnographic recording that include 19 EEG channels were obtained from 5 normal healthy young controls.
After sleep staging, for each subject, 5 different CAP A phase subtype epochs were selected, which served for subsequent analysis. Following
the analysis of power spectra calculated on the C4 channel by means of the fast Fourier transform, two different frequency bands were
detected: 0.252.5 and 712 Hz, representing the frequency peak in the profiles of the different CAP subtypes. All the subsequent analyses
were performed on these two bands. Scalp topographic color mapping was carried out using the data from all the 19 EEG channels recorded,
and by means of the 4-nearest neighbor algorithm. Individual average maps were obtained for both frequency bands. Finally, we used the low
resolution brain electromagnetic tomography (LORETA) functional imaging for the source analysis of the two EEG frequency components
of CAP A phases.
Results: The quantitative spectral analysis of the different A phase subtypes shows the existence of two distinct spectral components
characterizing CAP subtypes A1 (0.252.5 Hz) and A3 (712 Hz). These two components coexist in CAP A2 subtypes. The topography of
these two components shows a clear prevalence over the anterior frontal regions for the 0.252.5 Hz band and over the parietaloccipital
areas for the 712 Hz band. Finally, the generators of the low-frequency component of CAP seemed to be localized mostly over the frontal
midline cortex; on the contrary, those of the high-frequency band involved both midline and hemispheric areas within the parietal and
occipital areas.
Conclusions: The results of this study confirm the presence of two fundamentally distinct frequency bands which are expressed
individually (A1 and A3) or in association (A2) in the different CAP A phase subtypes. The analysis of scalp distribution maps indicates that
the two frequency components recognized are distributed over clearly different areas of the scalp. Moreover, the LORETA analysis indicates
that also the probable cortical generators of these two frequency bands are different and well separated and distinct.
q 2004 Elsevier B.V. All rights reserved.

Keywords: Cyclic alternating pattern (CAP); Scalp EEG topography; Cortical generators; EEG bands; Low resolution brain electromagnetic tomography

1. Introduction during NREM sleep with a specific form described as

Sleep EEG contains a large amount of phasic events,
such as sleep spindles, K-complexes, vertex waves, short-
lasting arousals, etc. These phasic events can occur
* Corresponding author. Tel.: C30-935-936111; fax: C39-935-653327.
E-mail address: rferri@oasi.en.it (R. Ferri).
1389-9457/$ - see front matter q 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.sleep.2004.06.010
cyclic alternating pattern or CAP [1,2], which consists
of transient arousal complexes (phase A) which interrupt
periodically the tonic theta/delta activities of NREM
sleep (phase B). Functionally, CAP is believed to
indicate a condition of sustained arousal instability
oscillating between a greater arousal level (phase A)
and a lesser arousal level (phase B). In this view,
30 R. Ferri et al. / Sleep Medicine 6 (2005) 2936
the absence of CAP might coincide with a condition of
arousal stability, characterized by the lack of arousal-
related phasic events and defined as non-CAP.
With the terms synchronization and desynchronization,
Terzano et al. [1,2] referred mostly to the frequency content
of CAP A phases; synchronization was expressed by the low
frequency components of K-complexes and delta bursts
during sleep stages 24, and desynchronization was
reflected by the high frequency components (in the alpha
and beta bands) in all sleep stages.
The different frequency content of CAP A phases is one
of the most important features allowing their visual
subclassification into three subtypes [3]:
Subtype A1. A phases in which EEG synchrony is the
predominant activity. In A1 phases the following
activities are included: intermittent alpha rhythm and
vertex sharp waves during stage 1, and sequences of
K-complexes or delta bursts in other NREM stages. If
present, EEG desynchronization occupies less than 20%
of the entire phase A. Subtype A1 is generally associated
with mild autonomic and somatomotor activity.
Subtype A2. A phases which contain a mixture of slow
and rapid EEG activities. In A2 phases the following
activities are included: K-alpha, arousal preceded by
slow wave synchronization and polyphasic bursts with
more than 20% but less than 50% of EEG desynchro-
nization. Subtype A2 is linked with a moderate increase
of muscle tone and/or cardio-respiratory rate.
Subtype A3. A phases with predominant EEG desyn-
chronization, including arousals, transient activation
phases or polyphasic bursts with more than 50% of
EEG desynchronization. Subtypes A3 are coupled with
an enhancement of muscle tone and/or cardiorespiratory
rate.
CAP rate is generally defined as the percentage of
NREM sleep or of each NREM sleep stage occupied by
CAP sequences [1,2]. This type of approach has been
conditioned by the common practice in sleep labs of
including only a limited number of derivations in the
polysomnographic recordings and by the internationally
accepted rules which only request one EEG channel [4],
even if additional EEG channels are required [3] for
CAP scoring. However, the significant advancements in
computer technology of the last years now allows the
recording and analysis of a large number of channels
with limited additional effort.
In this paper we present the results of the analysis of the
CAP A phases in a group of healthy normal controls, by
means of a computerized approach in order to define
quantitatively their spectrum content, their scalp topography
and their probable cortical generators and relative inter-
connecting pathways, by using data from sleep polygraphic
recordings that include a large number of scalp EEG
channels.
2. Subjects and methods
2.1. Subjects and polysomnographic recording
Five normal healthy controls (4 females and 1 male, aged
2032 years) were included in this study. They all had
regular life routine, did not smoke and did not take any
alcohol drink in the 3 days preceding the study.
All subjects underwent one overnight polysomnographic
recording, after one adaptation night, which comprised
EOG (2 channels), EEG (19 channels, electrodes placed
according to the 1020 International System referred to
linked earlobes, see Fig. 1 for channel details), EMG of the
submentalis muscle and ECG (see Fig. 1). Recordings were
carried out using a Brain Quick Micromed System 98
recording machine and signals were sampled at 256 Hz and
stored on hard disk in European data format (EDF) [5] for
further analysis. EEG signals, in particular, were digitally
band-pass filtered at 0.1120 Hz, 12-bit A/D precision.
Sleep stages were scored following standard criteria [4]
on 30-s epochs; subsequently, CAP and NCAP sequences
were visually detected in each recording, during S2 and
SWS, according to the rules defined by Terzano et al. [3],
and manually marked on screen by means of the sleep
analysis software Hypnolab 1.2 (SWS Soft, Italy). For each
subject, five different CAP A phase subtype epochs were
selected, with a minimum duration of 4 s. This number was
chosen because the number of A3 CAP subtypes (without
evident movement and/or muscle artifacts) usually found in
normal sleepers is low; this would have not allowed the
collection of a bigger number of epochs. Consequently, also
the number of the other subtypes was kept the same, in order
to have a reliable comparison between them.
2.2. Spectral analysis of CAP a phase subtypes
Power spectra were calculated for the C4 channel using
the sleep analysis software Hypnolab 1.2 (SWS Soft, Italy),
after Welch windowing wn Z 1K n1=2N K 1=1=2
N C 12 ; in order to minimize the truncation error and
reduce spectral leakage by suppressing sidelobes [6], by
means of the Fast Fourier Transform [7], on the first 4-s
artifact-free epochs from the different CAP A phase
subtypes selected before. Individual averages were
obtained, followed by group averages; these final averages
(and SD) were plotted for frequencies between 0.25 and
32.0 Hz. After a first analysis of the power spectra profiles,
two different frequency bands were detected: 0.252.5 and
712 Hz, representing the low-frequency peak in the
profiles of the A1 and A2 CAP subtypes and the high-
frequency peak observed in the spectral profiles of the A2
and A3 subtypes, respectively (see Section 3). All the
subsequent analyses were performed on these two bands
calculated, respectively, from the A1 and A3 CAP subtypes
previously selected.
R. Ferri et al. / Sleep Medicine 6 (2005) 2936 31

2.3. Scalp topographic mapping of the spectral

components of CAP

Scalp topographic mapping was carried out, for the

two bands described above, using the data from all the
19 EEG channels recorded and using the 4-nearest

Fig. 1. Example of our multichannel polygraphic recordings in which a typical sequence of CAP A phases is evident, marked by boxes; 60 s of recording are shown.
neighbor algorithm. With this approach, for every point
of the output map, a weighted average of the four closest
recorded data points (EEG channels) is performed; the
weighting factor used was 1/d2i , where di is the distance
from the point being interpolated to the data point i [8].
Different levels of spectral density were represented on a
color scale adjusted for each subject. Maps consisted of
128!128 pixels on the axial and sagittal planes (right
and left side views) and were obtained by means of the
sleep analysis software Hypnolab 1.2 (SWS Soft, Italy).
Also in this case, individual average maps were obtained
for both frequency bands.

2.4. Cortical source analysis of CAP by low resolution

brain electromagnetic tomography

We used the so-called low resolution brain electro-

magnetic tomography (LORETA) functional imaging for
the source analysis of the two EEG frequency com-
ponents of CAP A phases, which has been extensively
tested with simulation paradigms [9,10] and is presently
used by several independent labs worldwide. LORETA
computes 3-D linear solutions (LORETA solutions) for
the EEG inverse problem within a three-shell spherical
head model including scalp, skull, and brain compart-
ments. The brain compartment is restricted to the cortical
gray matter/hippocampus and is coregistered to the
Talairach probability brain atlas, digitized at the Brain
Imaging Center of the Montreal Neurologic Institute [11].
This compartment includes 2394 voxels (7-mm resol-
ution), each voxel containing an equivalent current
dipole. LORETA solutions consist of voxel current
density values able to predict EEG spectral power
density at scalp electrodes.
Solutions of the EEG inverse problem are under-
determined and ill-conditioned when the number of
spatial samples (electrodes) is lower than that of the
unknowns (current density at each voxel). To account for
that, cortical LORETA solutions predicting scalp
EEG spectral power density are regularized to estimate
distributed rather than punctual EEG sources [9,10]. As a
result, the spatial resolution of regularized
LORETA solutions is much lower than that of SPECT
or PET.
Individual average representations of the LORETA
solution were obtained for both frequency bands
considered in this study (from A1 and A3 CAP subtypes)
which were used to compute group grand averages.
32 R. Ferri et al. / Sleep Medicine 6 (2005) 2936
3. Results
Fig. 1 shows an example of the appearance of a typical
CAP sequence in our polysomnographic recordings. In this
example, the different A CAP subtypes are shown in boxes;
the gaps between the different boxes correspond to the B
phase of the CAP cycles.
Fig. 2 shows the results the spectral analysis of the
different CAP A subtypes. These spectra are the group grand
averages and also the SD of the mean is shown (dotted
lines). This analysis clearly shows that, as expected from the
visual detection rules, the spectrum of the A1 subtype is
characterized by a prominent peak in the low delta range,
Fig. 2. Results of the power spectrum analysis. Spectra were obtained from
the C4 channel.
between 0.25 and 2.5 Hz. The same peak is also evident in
the spectrum obtained from the A2 subtype, with slightly
smaller height than in the spectrum of the A1 subtype. In
this case, however, also more power is evident in the
frequencies ranging approximately between 7 and 12 Hz.
The last spectral profile, obtained from the A3 subtype, is
dominated by the presence of an evident peak in the alpha
band, at around 9 Hz, followed by some power expressed in
the high alpha and low beta bands (up to 1213 Hz).
Fig. 3 displays the results of the topographic scalp
mapping of the two frequency bands detected above; in
particular, the low-frequency band (0.252.5 Hz) shows,
in the different subjects, a clear prevalence over the
anterior frontal regions, mostly over the midline and
symmetrically spreading over the two hemispheres. On
the contrary, the high-frequency band (712 Hz) involves
mostly the parietaloccipital areas; also in this case, a
symmetrical distribution is evident with the peak over the
midline. In this band, a more variable distribution than
that of the slow-frequency band is evident in the different
subjects.
Fig. 4 shows, on the top, the results (grand average) of this
analysis at 1 Hz, representative of the probable cortical
generators of the slow-frequency component of CAP
obtained in our normal subjects. With this approach, the
generators of the low-frequency component of CAP seem to
be localized mostly over the frontal midline cortex. The same
figure shows, on the bottom, the results of LORETA at 9 Hz,
representative of the probable cortical generators of the high-
frequency component of CAP, which involve both midline
and hemispheric areas within the parietal and occipital areas.
Finally also CAP A2 phase subtypes were individually
analyzed by means of the LORETA method; however, in
this case we subdivided them into two parts: the first
(Fig. 5(a)) contained generally low-frequency components
with a cortical distribution not distinguishable from that
of the low-frequency component of CAP A1 phases, and
the last part (Fig. 5(b)) with high-frequency content and
probable cortical generators not distinguishable from
those of the high-frequency component of CAP A3
phases.
4. Discussion
The results of this study confirm the visual rules for CAP
detection based on the presence of two fundamentally
distinct frequency bands [3] which are expressed individu-
ally (A1 and A3) or in association (A2) with the different
CAP A phase subtypes. In addition, from the quantitative
spectral point of view, these bands seem to be well
separated, without overlap between them. In this study we
have directed our attention to the computerized analysis of
these two frequency components.
The analysis of scalp distribution maps indicate that
the two frequency components recognized by the visual
 R. Ferri et al. / Sleep Medicine 6 (2005) 2936
Fig. 3. Scalp topographic mapping of the two frequency components of CAP considered in this study.
and spectral analysis are distinct and map over clearly
different areas of the scalp. Moreover, the LORETA
analysis indicates that these two frequency bands are
different not only from the frequency and scalp distribution
points of view but also their probable cortical generators are
well separated and distinct.
This is also evident in the analysis of CAP A2 phase
subtypes (Fig. 5), which show in their first part a cortical
distribution not distinguishable from that of the low-
frequency component of CAP (characteristic of A1
subtypes) and in their last part probable cortical generators
not distinguishable from those of the high-frequency
component of CAP (typically evident in A3 subtypes). It
is important to emphasize that this association is not casual
and that the low-frequency component of CAP precedes the
occurrence of the high-frequency rhythms in the vast
majority of CAP A2 phases [12,13].
It is important to underline that our results, even if
obtained by averaging a relatively small number of epochs,
seem to be in agreement with those obtained by other
authors who did not point their attention to the CAP phases
33
but analyzed phasic activities, most of which were surely
part of CAP sequences, such as the study by Happe et al.
[14], who focused their attention on spontaneous
K-complexes and delta waves; also these authors found
that the power of the slow component of these waves shows
a peak over the medio-frontal regions. Interestingly, these
authors found that the delta frequency components of
K-complexes and delta waves are unaffected by spindles;
spindles are not considered in the scoring of CAP [3]. In
addition, evoked K-complexes seem to map over the
midfrontal areas [15] and the same frontal predominance
of the delta band was reported in another study by Finelli et
al. [16] in which spectra were obtained by averaging a large
number of epochs.
In this respect, one must consider that averaging EEG
epochs is potentially able to induce a serious distortion of
the results because sleep is rich in phasic events which, by
definition, have a short duration and their contribution to
average results can be very variable, depending not only on
the sleep stage but also on the presence or absence of CAP
sequences. Moreover, in the vast majority of studies on
34 R. Ferri et al. / Sleep Medicine 6 (2005) 2936

Fig. 4. Low resolution brain electromagnetic tomography (LORETA) functional imaging for the cortical source analysis of the two EEG frequency components
of CAP, at 1 and 9 Hz.

sleep many A3 phases (which often contain movement of epochs might have affected seriously the results obtained
artifacts and muscle potential contamination) have been by Coatanhay et al. [17] who applied LORETA to the study
considered as artifacts and excluded from the analysis. We of sleep potentials but averaged a large number of epochs
think that the problems induced by averaging a large amount from REM, stage 2 and slow wave sleep; they indicated

Fig. 5. Low resolution brain electromagnetic tomography (LORETA) functional imaging for the source analysis of the two EEG frequency components of
CAP, at 1 Hz (indicated as A2 (a) and 9 Hz (indicated as A2 (b). In particular, A2 a was obtained from the first 2 s of activity of the A2 CAP phase shown on the
left and A2 b from the following 2 s.
 R. Ferri et al. / Sleep Medicine 6 (2005) 2936
the left inferior temporal lobe as the cortical generator of
delta waves during slow-wave sleep and the left inferior
frontal cortex as the cortical generator of delta waves during
sleep stage 2. However, it is difficult to understand why only
these authors obtained lateralized source generators for
potentials which have repeatedly been described as bilateral
and symmetrical in the literature [1416].
The significance of the slow components of CAP phase A
subtypes, as EEG expression of cerebral activation during
sleep, has been a point of controversy for long time. This is
because the slow EEG components of sleep, such as
K-complexes and transient delta activities, were excluded
by the criteria of detection of arousals during sleep, defined
by the American Sleep Disorders Association [18], unless
these patterns are associated with an EEG frequency shift
towards theta, alpha or beta rhythms; the latter were
considered to be the true markers of cortical arousal.
Clinical evidence suggests that the definition of arousal
should be broadened [19,20] and include also very short
slow EEG phasic events [21].
From another point of view, the slow and rapid EEG
components can be both viewed as a sign of activation at
different levels of the brain during sleep and these different
patterns can allow to distinguish between cortical, sub-
cortical and autonomic arousals [22]. In fact, the A1
subtypes compose more than 90% of all the A phases
collected in the descending branches of the hypnogram and
in the troughs, while the A2 and A3 subtypes are the major
representatives of the A phases occurring in the ascending
branches. Thus, within the dynamic organization of sleep,
the non-random distribution of CAP sequences, with their
succession of slow (subtypes A1) and rapid (subtypes A2
and A3) EEG shifts, seem to be responsible for sculpturing
EEG synchrony under the driving and alternating forces of
NREM and REM sleep [23].
Synchronizing mechanisms during NREM sleep are
usually thought to be subserved by thalamocortical path-
ways [24,25]; based on our results, the CAP slow
component might be considered as the cortical expression
of this corticalsubcortical interaction. On the other hand,
the high-frequency component of CAP seems to be the
expression of the activity located at the level of the same
structures thought to generate alpha waves during wakeful-
ness, within the cerebral cortex at the level of the pyramidal
neurons in layers IV and V [26] with a system of surface-
parallel intracortical neurons involved in its spread [27,28].
The evidently nonrandom association between the two
different frequency components of CAP, the generators of
which are located in clearly different structures, seem to
indicate that these structures might be in functional
intercoupling allowing them to organize the occurrence of
the two different frequency components of CAP in a
coordinated way [22,23].
The superior longitudinal (arcuate) fasciculus is the
largest of the fiber bundles [29,30] that, together with
the uncinate and fronto-occipital fasciculi, constitute
35
the longitudinal association fiber system that connects
each frontal lobe with its respective hemisphere [31,32]. We
think that these important bundles might be of crucial
importance for the integration of the activity of frontal
generators and parieto-occipital generators of CAP. In this
respect, it should be noted that through these bundles
nervous signals can be transferred from the frontal lobes to
the parietal and occipital areas (and vice versa) in as few as
15 ms or less [33]; this speed might account for the probable
reciprocal inhibition during A1 and A3 CAP phases, which
might be so fast that only one frequency component is
evident on the scalp. However, these considerations are
highly speculative and more research is needed in order to
clarify the complex interactions between sleep and wake-
fulness neurophysiological mechanisms, probably taking
place during CAP in human sleep.
References
[1] Terzano MG, Mancia D, Salati MR, et al. The cyclic alternating
pattern as a physiologic component of normal NREM sleep. Sleep
1985;8:13745.
[2] Terzano MG, Parrino L, Spaggiari MC. The cyclic alternating pattern
sequences in the dynamic organization of sleep. Electroencephalogr
Clin Neurophysiol 1988;69:43747.
[3] Terzano MG, Parrino L, Smerieri A, et al. Consensus report. Atlas,
rules, and recording techniques for the scoring of cyclic alternating
pattern (CAP) in human sleep. Sleep Med 2001;2:53753.
[4] Rechtschaffen A, Kales A. A manual of standardized terminology,
techniques and scoring system of sleep stages of human subjects.
Washington Public Health Service, US Government Printing Office;
1968.
[5] Kemp B, Varri A, Rosa AC, et al. A simple format for exchange of
digitized polygraphic recordings. Electroencephalogr Clin Neurophy-
siol 1992;82:3913.
[6] Press WH, Flannery BP, Teukolsky SA, Vetterling WT. Numerical
recipes. The art of scientific computing. Cambridge: Press Syndicate
of the University of Cambridge; 1989.
[7] Cooley JW, Tukey OW. An algorithm for the machine calculation of
complex Fourier series. Math Comput 1965;19:297301.
[8] Shepard DS. A two-dimensional interpolation function for irregularly
spaced data Proceedings of 23rd National Conference, Association for
Computing Machinery. Princeton, NJ: Brandon/Systems Press; 1968.
[9] Pascual-Marqui RD, Michel CM, Lehmann D. Low resolution
electromagnetic tomography: a new method for localizing electrical
activity in the brain. Int J Psychophysiol 1994;18:4965.
[10] Pascual-Marqui RD, Lehmann D, Koenig T, et al. Low resolution
brain electromagnetic tomography (LORETA) functional imaging in
acute, neuroleptic-naive, first-episode, productive schizophrenia.
Psychiatry Res 1999;90:16979.
[11] Talairach J, Tournoux P. Co-planar stereotaxic atlas of the human
brain: three-dimensional proportional system. Stuttgart: Georg
Thieme; 1988.
[12] Terzano MG, Parrino L, Rosa A, et al. CAP and arousals in the
structural development of sleep: an integrative perspective. Sleep Med
2002;3:2219.
[13] De Carli F, Nobili L, Beelke M, et al. Quantitative analysis of sleep
EEG microstructure in the time frequency domain. Brain Res Bull
2004 in press.
[14] Happe S, Anderer P, Gruber G, et al. Scalp topography of the
spontaneous K-complex and of delta-waves in human sleep. Brain
Topography 2002;15:439.
36 R. Ferri et al. / Sleep Medicine 6 (2005) 2936
[15] Gora J, Colrain IM, Trinder J. The investigation of K-complex and
vertex sharp wave activity in response to mid-inspiratory occlusions
and complete obstructions to breathing during NREM sleep. Sleep
2001;24:819.
[16] Finelli LA, Borbely AA, Achermann P. Functional topography of the
human nonREM sleep electroencephlogram. Eur J Neurosci 2001;13:
228290.
[17] Coatanhay A, Soufflet L, Staner L, Boeijinga P. EEG source identi-
fication: frequency analysis during sleep. C R Biol 2002;325:27382.
[18] American Sleep Disorders Association. Arousals: scoring rules and
examples. A preliminary report from the sleep disorders atlas task
force of the American Sleep Disorders Association. Sleep 1992;15:
17484.
[19] Martin SE, Wraith PK, Deary IJ, Douglas NJ. The effect of nonvisible
sleep fragmentation on daytime function. Am J Respir Crit Care Med
1997;155:1596601.
[20] Pitson DJ, Stradling JR. Autonomic markers of arousal during sleep in
patients undergoing investigation for obstructive sleep apnoea, their
relationship to EEG arousals, respiratory events and subjective
sleepiness. J Sleep Res 1998;7:539.
[21] Thomas RJ. Arousals in sleep-disordered breathing: patterns and
implications. Sleep 2003;26:10427.
[22] Halasz P, Terzano M, Parrino L, Bodizs R. The nature of arousal in
sleep. J Sleep Res 2004;13:123.
[23] Terzano MG, Parrino L, Boselli M, et al. CAP components and EEG
synchronization in the first 3 sleep cycles. Clin Neurophysiol 2000;
111:28390.
[24] Steriade M, Contreras D, Curro-Rossi R, Nunez A. The slow (!1 Hz)
oscillation in reticular thalami and thalamocortical neurons: scenario
of sleep rhythms generation in interacting thalamic and neocortical
networks. J Neurosci 1993;13:328499.
[25] Contreras D, Steriade M. Cellular basis of EEG slow rhythms: a study
of dynamic corticothalamic relationships. J Neurosci 1995;15:
60422.
[26] Lopes da Silva FH, Storm van Leeuwen W. The cortical source of
alpha rhythm. Neurosci Lett 1977;6:23741.
[27] Lopes da Silva FH, Storm van Leeuwen W. The cortical alpha rhythm
in dog: depth and surface prophile of phase. In: Brazier MAB,
Petsche H, editors. Architecture of the cerebral cortex. IBRO
Monograph Series. New York: Raven Press; 1978. p. 331933.
[28] Lopes da Silva FH, Vos JE, Mooibroek J, van Rotterdam A. Relative
contribution of intracortical and thalamo-cortical processes in the
generation of alpha rhythms, revealed by partial coherence analysis.
Electroencephalogr Clin Neurophysiol 1980;50:44956.
[29] Klingler J, Gloor P. The connections of the amygdala and of the
anterior temporal cortex in the human brain. J Comp Neurol 1960;
115:33369.
[30] Williams PL, Worwick R. Grays anatomy. Edinburgh/London:
Churchill; 1980.
[31] Gloor P. The temporal lobe and the limbic system. New York: Oxford
University Press; 1997.
[32] Kiernan JA. Barrs the human nervous system: an anatomical view
point. Philadelphia: LippincotRaven; 1998.
[33] Zappoli R. Permanent or transitory effects on neurocognitive
components of the CNV complex induced by brain dysfunctions,
lesions and ablations in humans. Int J Psychophysiol 2003;48:
189220.