Brain, Vol. 122, No. 4, 687-707,
April 1999
© 1999 Oxford University Press
The role of the hippocampus in auditory processing studied by event-related electric potentials and magnetic fields in epilepsy patients before and after temporal lobectomy
1 Departments of Brain Pathophysiology, 2 Neurosurgery and 3 Neurology, Kyoto University Faculty of Medicine, Shogoin, Sakyo-ku, Kyoto, 606-8507, Japan
Correspondence to:
Hiroshi Shibasaki, MD, Department of Brain Pathophysiology, Kyoto University Faculty of Medicine, Shogoin, Sakyo-ku, Kyoto, 606-8507, Japan E-mail: shib{at}kuhp.kyoto-u.ac.jp
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsResultsDiscussionReferences |
|---|
To clarify the relationship between the hippocampus and the event-related responses in auditory information processing, we recorded event-related potentials (ERPs) and event-related magnetic fields (ERFs) associated with the auditory oddball paradigm in 12 patients with temporal lobe epilepsy before and after surgical treatment, and in eight age-matched healthy volunteers. Lesions in the patients were hippocampal sclerosis (8), cyst (2), cavernoma (1) and calcified arteriovenous malformation (1), all in the unilateral temporal lobe. Standard temporal lobectomy (8), selective amygdalohippocampectomy (2), selective hippocampectomy (1) and inferior lateral temporal resection (1) were carried out. ERPs were recorded in nine patients before surgery, in all 12 patients after surgery, and in all normal subjects. P300 was maximal at Pz in the patients both before and after surgery, and in normal subjects. The peak latency and amplitude of P300 measured at Pz in the patients either before or after surgery did not differ significantly from those in normal subjects. After surgery, only the amplitude of P300 over the anterior and mid-temporal area on the resected side was attenuated, while it was symmetric before surgery regardless of the side of epileptogenic focus. ERFs were recorded in three patients before surgery and in six normal subjects by using a whole-head neuromagnetometer. ERFs in response to the target stimuli at a latency of ~400 ms were recognized at the anterior, middle and posterior lateral channels on each hemisphere (M400). The latency and dipole moments for M400 did not differ significantly between the patients before surgery and the normal subjects. As a result of analysis using the time-varying multidipole model, three dipoles for M400 were estimated in two patients in whom ERFs were available before surgery for the analysis, and in normal subjects: mesial temporal area, superior temporal area and inferior parietal area on each hemisphere. After surgery, in four out of six patients in whom ERFs were recordable, M400 at the anterior temporal channels on the resected side disappeared, and the activity in the affected mesial temporal area was lost. In one patient who underwent inferior lateral temporal resection, M400 waveforms and its sources were preserved in all regions. There were no significant differences in the latency and dipole moments of the unaffected source of M400 before versus after surgery. These results suggest that the hippocampus contributes to the scalp-recorded P300 only at the corresponding anterior temporal region, and does not influence its general waveform and predominant distribution over the scalp.
hippocampus; temporal lobectomy; event-related potentials; event-related magnetic fields; magnetoencephalography
ANOVA = analysis of variance; BNE = balanced non-cephalic electrode; ECDs = equivalent current dipoles; EOG = electro-oculogram; ERFs = event-related magnetic fields; ERPs = event-related potentials; ILTR = inferior lateral temporal resection; MEG = magnetoencephalography; SAH = selective amygdalohippocampectomy; SH = selective hippocampectomy; STL = standard temporal lobectomy
|
Introduction |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
It has been suggested that event-related potentials (ERPs), especially P300, reflect neural activity related to attention and memory updating of discrete events (Vaughan and Ritter, 1970
; N. K. Squires et al., 1975; Courchesne et al., 1977; K. C. Squires et al., 1977; Donchin et al., 1978; Barrett et al., 1987; Donchin and Coles, 1988; Fabiani et al., 1987; Picton and Hillyard, 1988; Neshige and Lüders, 1992; Suwazono et al., 1994; Kanda et al., 1996; Nishitani et al., 1996). Regarding the generator source of P300, several intracranial depth recordings for localizing the epileptogenic foci in patients with intractable partial epilepsy provided some evidence supporting P300 being generated at least from the hippocampus and amygdala (Halgren et al., 1980, 1995; N. K. Squires et al., 1983; Johnson and Fedio, 1986 Johnson and Fedio, 1987; McCarthy and Wood, 1987; Stapleton and Halgren, 1987; Stapleton et al., 1987; Halgren, 1988; Johnson, 1989a, b; McCarthy et al., 1989). In the analysis of the intracerebral potentials, the P300 wave was frequently abnormal when recorded from the mesial temporal area containing the sclerotic hippocampus (Meador et al., 1987; Wood et al., 1988; Puce et al., 1989). There were some studies of the scalp-recorded P300 in patients with lesions in the mesial temporal area. Rugg et al. (1991a) reported that P300 was symmetric and of normal amplitude in a patient with severe verbal memory deficit due to low grade glioma involving the full extent of the left mesial temporal area. Polich and Squire (1993), by recording P300 in patients with amnesia due to bilateral hippocampal lesions including the CA1 region, reported that the hippocampal formation did not significantly contribute to the scalp-recorded P300. Honda et al. (1996) found P300 abnormalities in patients with selective impairment of recent memory, and concluded that the potential generated in the mesial temporal area, by itself, is not always reflected as a component of the scalp-recorded P300. Furthermore, the scalp-recorded P300 was normal in epileptic patients even after the unilateral temporal lobectomy (Stapleton and Halgren, 1987; Stapleton et al., 1987; Johnson, 1988). O'Donnell et al. (1993) showed, by using the Brain Electrical Source Analysis (BESA) (Scherg and Von Cramon, 1986; Scherg, 1992), that the dipoles of P300 in the mesial temporal area were affected in patients with amnestic syndrome. These studies suggest that the scalp-recorded P300 may represent the summation of the activities arising from a number of different generators, and that the hippocampus is not the only source of the scalp-recorded P300.
As EEG is strongly affected by the different electric conductivity of the structures covering the brain surface, especially the cerebrospinal fluid and skull, it is difficult to evaluate EEG signals accurately in patients, especially after craniotomy. On the contrary, since the magnetic field generated in the brain is influenced much less by those surrounding structures compared with electrical potentials (Barth et al., 1986), magnetoencephalography (MEG) can identify the neural source more accurately than EEG in the patients, even after craniotomy. For the possible sources of event-related magnetic fields (ERFs) recorded in an auditory oddball paradigm, corresponding to P300, two regions have been suggested so far; the mesial temporal area, possibly the hippocampus (Okada et al., 1983; Lewine et al., 1990), and the superior temporal area (Gordon et al., 1987). Rogers et al. (1991), also employing the auditory oddball paradigm, concluded that the responses to the target stimuli arose in deep cerebral regions and propagated to the auditory cortex with time. Okada et al. (1983) recorded ERFs in a visual oddball paradigm, and identified the sources for their P3m in the hippocampus. However, these studies recorded responses only from a restricted area of the head with limited MEG channels and, furthermore, utilized a single dipole model. Recently, two groups reported the neural correlates of auditory processing by using a whole-head MEG (Tesche et al., 1996; Nishitani et al., 1998). Tesche et al. (1996) detected the neural sources in the mesial temporal area in a high stimulus rate auditory detection task. We demonstrated three generator sources for the ERF on each hemisphere in the auditory oddball paradigm; the mesial temporal area, the superior temporal area and the inferior parietal area (Nishitani et al., 1998).
This is the first combined study of ERPs and ERFs recorded by whole-head MEG in patients with temporal lobe epilepsy before and after surgical treatment. By using MEG, which is a clinically useful method of estimating brain activities non-invasively without being influenced by changes in the structures surrounding the cortical surface, the results obtained clarified a relatively limited role for the hippocampus in the generation of the scalp-recorded P300.
|
Methods |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Subjects
Data were obtained from 12 patients (four males and eight females; age 19–47 years, mean 29 years; all right-handed) with the diagnosis of medically intractable temporal lobe epilepsy according to the criteria of International Classification of Epilepsies and Epileptic Syndromes (Commission on Classification and Terminology of the International League Against Epilepsy, 1989
). Clinical profiles of each patient are given in Table 1. Age at seizure onset was 2–20 years (mean 10.3 years), and the duration of clinical course after the first seizure was 7–40 years (22.2 years). All patients were treated with some anticonvulsants (one or various combinations of carbamazepine, diazepam, phenobarbital, phenytoin and valproic acid) before surgery, which were left unchanged after surgery. Ictal EEG showed focal epileptiform discharges starting from anterior temporal, mid-temporal or frontopolar channels over the right or left hemisphere. Head MRI revealed arteriovenous malformation, cavernoma, calcification and a cystic lesion in one patient each, and sclerosis or atrophy in seven patients, all in the mesial temporal area. In the remaining patient (Patient 12), a cystic lesion was seen in the inferior lateral temporal area. On the intracarotid sodium amobarbital (Wada) test (Wada and Rasmussen, 1960; Powell et al., 1987), the speech dominance was found on the left hemisphere in five patients, on the right in three and bilateral in one, and the memory dominance was found on the left hemisphere in five patients, on the right in one and bilateral in three (Table 1). In one patient (Patient 11), the Wada test could not be examined because of the patient's poor performance. With regard to surgical treatment, selective amygdalohippocampectomy (SAH) was performed in two patients, selective hippocampectomy (SH) and inferior lateral temporal resection (ILTR) in one each, and standard temporal lobectomy (STL) in the remaining eight patients. In STL, the superior temporal gyrus was preserved in all patients. SH resected the anterior 2–3 cm of the hippocampus. SAH consisted of total resection of the amygdala and removal of the anterior hippocampus (2–3 cm). ILTR included the partial resection of the inferior temporal and fusiform gyri.
|
For the control study, eight healthy volunteers (six males and two females; age 21–37 years, mean 28 years; seven right-handed and one left-handed) participated in the recording of ERPs, and six of them in the recording of ERFs. The ERF data of these normal subjects were reported previously for another purpose (Nishitani et al., 1998
). Informed consent was obtained from each patient after full explanation concerning the purpose of this study. This research was done in accordance with the guidelines of the Committee of Medical Ethics, Graduate School of Medicine and Faculty of Medicine, Kyoto University.
Experimental paradigm
This study was carried out using the same paradigm as that employed for our previous study (Nishitani et al., 1998). The subjects were seated in an arm-chair in a quiet room for the ERP recording. For the ERF recording, they were seated in a chair made of glass fibre in a quiet and magnetically shielded room, and their head was positioned in a helmet-shaped dewar and closely attached against its inner vault. They were requested to keep their eyes open, to fix on a target mark placed 2 m in front of them and to keep still during the measurements in order to eliminate blinks and slow eye movements.
The standard auditory oddball paradigm was employed for the stimulus sequence, in which the tones of either 1 kHz (non-target) or 2 kHz (target) were delivered binaurally by a sound stimulator with the probability of 80% for the non-target and 20% for the target stimuli. The duration of each tone was 50 ms, consisting of 10 ms of rise/fall and 30 ms of plateau. The order of tone presentation was randomized, and the rate of stimulus presentation was once per 2.2 s on average (2.0–2.4 s). For the ERP recording, the tones were presented through the headphone. For the ERF recording, in order to eliminate large magnetic artefacts, the tones were delivered from small speakers placed 5 cm away from the dewar of the neuromagnetometer, and were led to the subjects through plastic tubes and ear pieces which fitted tightly into the external ear canal. The intensity of both tones was adjusted to 85 dBSPL on the headphone for the ERP recording and at the exit of the plastic tube for the ERF recording. As the response task, the subjects were requested to count the number of the presented target stimuli silently and to report their total number at the end of each session.
Data acquisition
ERP recording
EEGs were recorded from 22 shallow cup electrodes placed on the scalp at FP1, FP2, F7, F3, Fz, F4, F8, T3, C3, Cz, C4, T4, T5, P3, Pz, P4, T6, O1, Oz and O2 according to the International 10–20 System, and additionally at T1 and T2. Those were affixed on the scalp with collodion, and the impedance of all electrodes was kept at <5 k
. In this recording, all electrodes were referenced to a balanced non-cephalic electrode (BNE) (Stephenson and Gibbs, 1951). A monopolar electro-oculogram (EOG) was monitored with an electrode placed below the right outer canthus referenced to the BNE. EEG and EOG signals were filtered by a bandpass of 0.05–100 Hz (–3 dB), and the sampling rate used for digital conversion was 500 Hz.
The responses were averaged on-line with respect to the onset of the target and non-target stimuli separately. The analysis window for the on-line averaging was from 100 ms before to 900 ms after each stimulus onset. Approximately 100 auditory stimuli, comprising both the target and non-target stimuli, were delivered in one session. In order to confirm the reproducibility of responses, at least three sessions of testing were carried out for each recording in each patient. Within each session, the response to the first target stimulus was excluded from the on-line average to avoid the startle effect. The epochs with amplitudes exceeding 100 µV for EEG and/or EOG were excluded automatically from the averaging. When any sessions were contaminated with excessive artefacts such as eye movements and noise in any of the channels or if the patient was found to be drowsy, those sessions were excluded from the following analysis, and additional sessions were carried out.
ERF recording
The brain's magnetic signals were measured with a helmet-shaped 122-channel neuromagnetometer, which used 61 pairs of two orthogonally arranged `figure of eight' planar first-order gradiometers placed at 61 measurement sites. This system measures the two orthogonal derivatives of the radial magnetic component (Ahonen et al. 1993), and typically detects the largest signal just above the corresponding generator source. Head position with respect to the sensor array was measured with head position indicator coils placed on the defined scalp sites. The positions of head position indicator coils in relation to anatomical landmarks of nasion and bilateral preauricular points, where oil tablets were placed for the MRI scanning, were measured with a 3D digitizer (Isotrak 3S1002, Polhelmus Navigation Sciences, Colchester, VT) to allow alignment of the MEG and MRI coordinate systems. At the beginning of every recording session, the magnetic signal produced by three head position indicator coils on the scalp was measured by the sensors to determine the head position with respect to the sensor array. Head MRIs were obtained with a one Tesla Siemens Magnetom(TM) system from all patients before and after surgery, and from all normal subjects.
During the MEG recording, EEG was recorded simultaneously from Ag/AgCl shallow cup electrodes placed at Fz, Cz and Pz in all subjects. The electrodes were affixed to the scalp with collodion and referenced to the linked ear lobe electrodes. The impedance of all electrodes was kept at <5 k. To monitor eye movements and blinks, a monopolar EOG from an electrode placed below the right outer canthus referenced to the linked ear lobe electrodes and a bipolar EOG from that electrode to the right supraorbital electrode were recorded. To confirm the subject's vigilance and the task performance, the waveforms of EEG, EOG and selected channels of MEG were monitored on a cathode-ray tube display throughout the recording. Short breaks were given to refresh the subjects between sessions. During the intermissions, they were requested to keep the same head position as far as possible. The recording bandpass was 0.03–100 Hz for MEG, and 0.07–120 Hz for EEG and EOG, and the sampling rate for digital conversion was 404 Hz for all of them. The conditions for the on-line averaging, analysis window and stimulus rate in the ERF recording were the same as those adopted for the ERP recording. The epochs containing MEG signals exceeding 1500 fT/cm and those containing EEG or EOG signals exceeding 150 µV were excluded automatically from the averaging.
Data analysis
Identification of waveforms
For both MEG and EEG, after confirming the reproducibility of the averaged waveforms among like sessions for each individual subject, group averages for each of the target and non-target stimuli were obtained, and digitally low-pass filtered at 40 Hz prior to further analysis.
The peak of the N100 component of the EEG response was determined visually at Cz with the peak latency ~100 ms after the stimulus onset, and that of the N1m component of the MEG response was also detected visually at the channels showing maximum magnetic deflection on each hemisphere around the same latency as that of N100.
The P300 component of ERPs was determined as a positive deflection of the EEG response to the target stimuli seen at Pz at the latency between 250 and 600 ms. The amplitudes of the scalp P300 were measured at all electrodes at the latency of P300. The magnetic deflections corresponding to P300 (M400) were detected during the same latency range in the MEG channels where the target response showed larger deflection compared with the non-target one. Amplitude measurement was done from the baseline which was determined by averaging the 100 ms segment before the stimulus onset for each channel of MEG and EEG.
The topographic maps at N100 and P300 peak latency were obtained for each subject by using the DP1100 Topography software (NEC-Sanei).
Source modelling from MEG
The sources of the magnetic fields were modelled as equivalent current dipoles (ECDs), whose location, orientation and current strength were estimated from the measured magnetic waveforms. To identify the source of ECD, the spherical head model whose centre best fitted the local curvature of the subject's brain surface was adopted, based on the individual MRI for each subject (Hämäläinen et al., 1993). The ECDs that best explained the most dominant signals were determined by the least-squares search by using 20–30 channels at the areas including the local maximum signals. For each subset of channels, ECDs were calculated for every 1 ms segment over the time period of 50 ms containing the peak latency of each component. Only the ECDs which fulfilled the following two criteria were adopted: (i) goodness of fit (g-value), which indicates how satisfactorily the model accounts for the measured field variance (Kaukoranta et al., 1986), above 80% at selected periods of time for each subset of channels, and (ii) the confidence volume (Hämäläinen et al., 1993) below 300 mm3. ECDs with the highest g-value and the smallest confidence volume were accepted for further analysis. Thereafter, the analysis was extended to the whole recording time and to all channels, based on the assumption that the sources were fixed in location and orientation but their strengths and directions changed as a function of time, as described below (Scherg and Von Cramon, 1986; Mosher et al., 1992; Scherg, 1992; Hämäläinen et al., 1993). The isocontour maps at the selected latency were constructed from the magnetic fields by using the minimum-norm estimate that best accounts for the measured data (Ilmoniemi, 1991).
First, the ECDs for the N1m in response to the target and non-target stimuli were determined separately on each hemisphere. For each of the target and non-target responses, the predicted waveforms throughout the analysis window over the whole head obtained by introducing the two-dipole model for N1m were superimposed on the measured waveforms. Secondly, focusing on M400 in the target waveforms, a subset of channels, which showed a significant difference between the measured waveforms and the predicted ones obtained by adopting the above two-dipole model for N1m, was identified visually on each hemisphere, and then the ECDs were calculated over the subset of channels thus selected on each hemisphere. If any discrepancy still remained between the measured waveforms and the predicted ones obtained by adopting the dipoles for N1m and those for M400, another subset of channels was selected over other areas for detecting additional ECDs for M400. The remaining, unexplained part of the measured signals was extracted with the analysis method discussed in our previous study (Nishitani et al., 1998), and a new ECD was identified from the remaining signals. Every time a new ECD was introduced, the predicted waveforms were calculated by adopting the fixed multidipole model. If there were magnetic signals unexplained by the predicted waveforms, the data were re-evaluated for more accurate estimation. The g-values were calculated over all 122 channels and over the entire time period, and compared among different models to find the best possible solution. This approach was shown to give reliable results (Hämäläinen et al., 1993) and has been used successfully in previous studies (Hari et al., 1993; Levänen et al., 1996; Nishitani et al., 1998). Finally, the estimated dipoles obtained through these procedures were superimposed on the subject's own MRI according to the alignment of the MEG and MRI coordinate system which was determined as described above.
Statistical analysis
An analysis of variance (ANOVA) was performed to evaluate the effects of groups (normal subjects, patients before and after surgery) on the latencies of the N100 and P300 component of ERPs and the effects of groups and sides (epileptogenic/resected side and intact side) on the amplitudes of N100 elicited by target stimuli. Amplitudes of P300 were normalized by subtracting the minimum values from each value and dividing the results by the difference between the maximum and minimum values at each channel (McCarthy and Wood, 1985). To test the significance of amplitude differences across groups, the absolute as well as normalized amplitude of P300 in response to the target stimuli was analysed by ANOVA with respect to the effect of groups. These statistical analyses were also done in the difference waveforms obtained from the target and non-target stimuli in the range of P300 latency.
ANOVA was carried out in order to clarify whether there was any significant difference in the amplitude of P300 between hemispheres in the normal subjects. Regarding the patients' data, the epileptogenic and the resected sides were standardized to the same side. Based on the normalized amplitude of P300 in the responses to target stimuli as well as in the difference waveforms, the interactions between groups and electrodes were evaluated among all electrodes, among frontal electrodes (F7, F3, F4 and F8), among central electrodes (T3, C3, C4 and T4) and among posterior electrodes (T5, P3, P4 andT6), among parasagittal electrodes on each hemisphere [(Fp1, F3, C3, P3 and O1) and (Fp2, F4, C4, P4 and O2)], and among lateral temporal electrodes on each hemisphere [(F7, T3 and T5) and (F8, T4 and T6)]. The interactions between groups and electrodes on either parasagittal or lateral temporal electrodes were also estimated between both hemispheres. These interactions were tested between the normal subjects and the patients before surgery, and between the patients before and after surgery.
The latency and dipole moments of M400 in the normal subjects were analysed by ANOVA with the effect of hemispheres. Kruskal–Wallis test for the statistical analysis of the latency and dipole moments of M400 was performed among the left mesial temporal areas of the normal subjects, the mesial temporal area on the intact side in the patients before surgery and that after surgery. The same test for the latency and dipole moments of M400 among groups was done in the superior temporal area and the inferior parietal area on both the epileptogenic/resected side and the intact side.
|
Results |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Before surgery, both ERPs and ERFs were recorded in three patients, and ERPs alone in six patients. The recordings were made 3 days to 6 months before surgery. After surgery (4 months to 1 year), both ERPs and ERFs were recorded in 10 patients, and only ERPs in the remaining two. In four out of 10 patients who had ERF records after surgery, the magnetic signals were contaminated with large noise. In two other patients (Patients 4 and 9), the early responses to the target and non-target stimuli were evoked before and after surgery, but the late responses showed no difference between the target and non-target stimuli before surgery in Patient 4 and after surgery in Patient 9. Consequently, the ERF data were available for the analysis of the late responses both before and after surgery in only two patients (Patients 1 and 3), and in three other patients (Patients 5, 6 and 12) only the data obtained after surgery were analysed. In the patients who underwent both ERP and ERF recordings either before or after surgery, ERP and ERF studies were carried out on separate days <1 week apart, under the same dose of medications for each individual patient. In none of the patients did clinical seizure occur during the ERP or ERF recordings.
Task performance
In all healthy volunteers as well as patients, the task of counting the number of target stimuli was performed with an accuracy of >98%. The total number of stimuli per block was 100 for ERP in all subjects, and varied from 100 to 120 for ERF. The number of target stimuli per block for both ERP and ERF ranged from 15 to 25. For each subject, four or five recording blocks were obtained, and three or four blocks were averaged for further analysis of the ERP and ERF data after confirming the reproducibility of the waveforms among like blocks. As a result, the mean numbers of responses to the target stimuli and to the non-target stimuli were 65 and 331, respectively, per subject.
ERPs
Early responses (N100)
Latencies.
In the normal subjects, the early responses were evoked equally by both the target and non-target stimuli, and were also evoked in all patients equally before and after surgery. The responses were clearly detected at Cz (N100) (Fig. 1A). There were no significant differences in the peak latency of N100 measured at Cz among groups (Table 2).
|
|
Amplitudes.
The scalp topography of N100 in each patient before surgery showed the symmetric distribution, with the maximum amplitude at Cz, which was similar to the N100 distribution of the normal subjects (Fig. 1C
). There were no significant differences in the amplitude at Cz among groups. In the patients before surgery, there was no difference in the amplitude of N100 between the two hemispheres with respect to the epileptogenic side. In the patients who underwent SAH, SH and STL, the amplitude of N100 at the operated side increased, while it did not change at the unoperated side (Fig. 2). The amplitudes of N100 at the anterior lateral channels showed a significant difference between the operated and the unoperated side [F(1,22) = 10.0, P < 0.01] (Fig. 3A). The scalp topography of N100 obtained after surgery was significantly different from that obtained before surgery (Fig. 2C). In all the patients who underwent SH and STL and in one patient who underwent SAH (Patient 1), N100 showed asymmetric distribution with larger negativity at the fronto-temporal region on the operated side (Fig. 2C). In the patient who underwent ILTR (Patient 12), there was no difference in the amplitude of N100 over the anterior temporal region between the two sides (Fig. 3A).
|
|
Late responses (P300)
Latencies.
In all normal subjects and in all patients both before and after surgery, only the target stimuli elicited a positive component between 300 and 425 ms after the stimulus onset, with the maximum amplitude at Pz (Figs 1 and 2
). There were no significant differences in the peak latency among groups (Table 2).
Amplitudes.
For both the original and normalized data, there were no significant differences in the amplitude of P300 at Pz among groups, either in the responses to target stimuli or in the difference waveforms.
In normal subjects, there was no significant difference in the amplitude of P300 between hemispheres. When the results for the patients before surgery were compared with the normal subjects, there were no significant groups x electrodes interactions on the amplitude of P300, either in the normalized responses to target stimuli or in the difference waveforms, in any of the electrode combinations.
The scalp topography of P300 obtained after surgery was significantly different from that obtained before surgery (Fig. 2D). When the data of the patients were compared before and after surgery, there were significant groups x electrodes interactions on the amplitude of P300 among the lateral temporal electrodes [F(5,114) = 3.10, P < 0.05 and F(5,114) = 2.69, P < 0.05] and among the frontal electrodes [F(3,76) = 3.40, P < 0.05 and F(3,76) = 2.75, P < 0.05], both in the normalized responses to target stimuli and in the difference waveforms. However, there were no significant groups x electrodes interactions among other electrodes. There was a significant difference in the amplitude of P300 only at the anterior lateral electrode on the resected side before versus after surgery, in the responses to target stimuli as well as in the difference waveforms, by Scheffe's F post hoc test [lateral electrodes, P < 0.05 and P < 0.01; frontal electrodes, P < 0.05 and P < 0.005].
For the results in patients before surgery, there was no difference in the amplitude of P300 at the anterior lateral electrodes between the epileptogenic and intact sides, either in the responses to target stimuli or in the difference waveforms. After surgery, for both the responses to target stimuli and the difference waveforms, there were significant amplitude differences at the anterior lateral electrodes between the resected and intact sides in all patients except for one who underwent ILTR (Patient 12) [SAH/SH, F(1,4) = 8.60, P < 0.05 and F(1,4) = 8.98, P < 0.05; and STL, F(1,14) = 4.84, P < 0.05 and F(1,14) = 5.94, P < 0.05]. Both in the responses to target stimuli and in the difference waveforms, there were also significant differences in the amplitude of P300 at the anterior temporal electrode on the epileptogenic/resected side before versus after surgery [SAH/SH, F(1,3) = 12.3, P < 0.05 and F(1,3) = 60.3, P < 0.05; and STL, F(1,12) = 6.06, P < 0.05 and F(1,12) = 11.3, P < 0.05] (Fig. 3B). The patient who had had ILTR showed the symmetric distribution, and there was no attenuation or magnification of P300 amplitude on the operated side.
The ERPs recorded simultaneously during the ERF recording also showed P300 with the maximum amplitude at Pz in the patients both before and after surgery.
ERFs
Early responses (N1m)
Magnetic deflections in the responses to the target and non-target stimuli in Patient 1 are illustrated along with the simultaneously recorded EEG responses in Fig. 4. The early responses were evoked equally by both the target and non-target stimuli in the bilateral temporal areas (N1m). Table 3 shows the peak latency and the signal amplitude of N1m for the target stimuli in the normal subjects, and in the patients before and after surgery.
|
|
There were no significant differences in the peak latency of N1m between the two hemispheres in each group. The signal amplitude of N1m did not differ between the two hemispheres in the normal subjects or between the epileptogenic and intact hemispheres in the patients before surgery. In the patients who underwent the resection of hippocampus, the signal amplitude of N1m did not differ significantly between the hemisphere on the resected side and the intact hemisphere, regardless of the type of resection.
In both of the two patients in whom ERFs were available before and after surgery (Patients 1 and 3), the magnetic fields of N1m for the target and non-target stimuli showed dipolar patterns on each hemisphere. Superimposition of these ECDs of N1m on the patient's own MRI indicated their location in the Heschl's gyrus of each temporal lobe (Fig. 5). The locations of ECDs for N1m and M400 in the patients before and after surgery and those in the normal subjects are shown in Table 4. ECDs for N1m in the patients both before and after surgery were located in similar areas to those for the normal subjects.
|
|
The simulated waveforms of MEG responses to the non-target stimuli obtained by adopting two dipoles of N1m showed no significant differences from the measured waveforms throughout the entire analysis time. For the responses to the target stimuli, however, the simulated waveforms obtained by adopting two dipoles of N1m explained the measured waveforms only around the latency of N1m. During the later time period of the target responses, there were remarkable differences between the measured and simulated waveforms, thus indicating that the two-dipole model for N1m did not explain the late responses (Fig. 6
). These findings were common to all normal subjects and to all patients both before and after surgery, and also corresponded to the results obtained from the normal subjects in our previous study (Nishitani et al., 1998).
|
Late responses (M400)
Latencies.
In all normal subjects and in the two patients in whom the ERF data before surgery were available for the analysis (Patients 1 and 3), the deflections peaking between 330 and 430 ms after the stimulus onset were observed only in response to the target stimuli over the anterior, middle and posterior lateral channels on each hemisphere (M400) (Figs 4 and 7
). In normal subjects, there were no significant differences in the M400 latency between the two hemispheres either at the anterior, middle or posterior lateral channels. For the epileptogenic side, there were also no differences in the M400 latency among the patients before surgery, those after surgery and the normal subjects (data from both hemispheres), as was also the case for the intact hemisphere. In the normal subjects, the dipole moments of M400 in the anterior lateral, middle lateral and posterior lateral areas were not significantly different between hemispheres (Table 5).
|
|
Dipole moments.
In four out of the six patients for whom the ERF data after surgery were available (Patients 1, 3, 5 and 6), the late deflections were detected at the middle and posterior lateral channels on both hemispheres, and at the anterior lateral channel on the intact side but not on the operated side (Fig. 7
). In one patient who underwent ILTR (Patient 12), the late deflections were preserved in the anterior, middle and posterior lateral channels on both hemispheres. For the intact hemisphere, there was no significant difference in the dipole moments of M400 in the mesial temporal area among the patients before surgery, those after surgery and the normal subjects. With regard to the dipole moments of M400 in the superior temporal area and the inferior parietal area, there were no significant differences among the patients before surgery, those after surgery and the normal subjects, and either on the epileptogenic side or on the intact hemisphere (Table 5).
In Patient 4 before surgery and in Patient 9 after surgery, the late deflections in the MEG response to the target stimuli were not recognized at any channel (Fig. 7), although they performed the task well and also P300 of normal waveform was evoked on EEG.
Source modelling for M400
At the peak latency of M400 for each subset of channels, the magnetic fields of M400 showed dipolar patterns over each hemisphere in the patients before surgery, and were similar to those of the normal subjects of the present as well as our previous series (Fig. 8) (Nishitani et al., 1998). The magnetic fields of M400 in the patients after SAH, SH and STL showed similar dipolar patterns over each hemisphere, except for the resected area. In the patient after ILTR, the magnetic fields of M400 showed the dipolar patterns for all subsets of channels over both hemispheres.
|
As described in the section on early responses (N1m), the simulated waveforms obtained by adopting the two-dipole model for N1m were significantly different from the measured responses to target stimuli for the later component. This finding indicated that other dipoles were necessary to explain the late responses of the ERF waveforms to target stimuli. By employing the source localization procedure in all the ERF waveforms, a subset of channels was first chosen over the middle lateral channels on each hemisphere where the largest differences in the late component between the measured and simulated waveforms were observed. Since the simulated waveforms obtained by adopting these two dipoles for M400 were significantly different from the measured waveforms at the anterior lateral channels in all ERFs, another subset of channels was selected over the anterior lateral channels on each hemisphere. Since the simulated waveforms obtained by adopting those four dipoles for M400 were still significantly different from the measured waveforms over the posterior lateral channels, an additional subset of channels was selected at the posterior lateral channels on each hemisphere. For the results of the three-step analysis procedure, in the two patients in whom M400 data were available before surgery (Patients 1 and 3), three ECDs for M400 were detected on each hemisphere; one each in the mesial temporal area, the superior temporal area and the inferior parietal area. This result was consistent with that obtained in the normal subjects of our previous series (Nishitani et al., 1998
). When the simulated waveforms, obtained by adopting each of the two-, four- and six-dipole models for M400 in Patient 3 before surgery, were superimposed on the measured waveforms in response to the target stimuli, the waveforms obtained by adopting the six-dipole model explained the measured waveforms sufficiently for the later components. The g-value for the six-dipole model for M400 improved significantly (92% at the latency of 308 ms), compared with that for two- and four-dipole models (71 and 82%, respectively) (Fig. 9A). Furthermore, in Patient 3 before surgery, there were no residual differences between the measured waveforms and the simulated ones obtained by adopting the eight-dipole model, consisting of two for N1m and six for M400, throughout the analysis time at any channels. Similar results were obtained in Patient 1 before surgery.
|
After surgery, ECDs for M400 were detected in the superior temporal area and the inferior parietal area bilaterally and in the mesial temporal area only on the intact side in four patients (Patients 1, 3, 5 and 6) (Fig. 7
). In these patients after surgery, the simulated waveforms obtained by adopting the five-dipole model fitted well to the measured waveforms for the later components. In Patient 12 who underwent the left ILTR, all three ECDs on each hemisphere were detected in similar regions to those seen in the normal studies (Nishitani et al., 1998), and the simulated waveforms obtained by adopting the six-dipole model fitted well to the measured waveforms for the later components. The g-value for the five-dipole model in Patients 1, 3, 5 and 6, and the six-dipole model in Patient 12 improved the value compared with the two- or four-dipole model; e.g. 71, 79 and 86% for the two-, four- and five-dipole models, respectively, in Patient 3 (Fig. 9B). The five- or six-dipole model explained the measured magnetic signals best. Dipoles obtained from the four patients who were examined before as well as after surgery were superimposed on each subject's own MRI (Fig. 10).
|
There were no significant differences in x-, y- and z-coordinates for those three ECDs for M400; in the mesial temporal area, the superior temporal area and the inferior parietal area on each hemisphere, among groups (Table 4
). ECDs for M400 were not affected by the changes of the structures covering the brain surface caused by surgical treatment. |
Discussion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
In order to clarify the relationship between the mesial temporal area and the scalp-recorded P300 in auditory information processing, we recorded ERPs and ERFs associated with the conventional auditory oddball paradigm in patients with medically intractable temporal lobe epilepsy before and after resection of a part of the temporal lobe. Before surgery, the scalp distribution of P300 was maximal at the midline vertex and symmetric, and there was no significant difference in its amplitude between the two hemispheres irrespective of the side of the epileptogenic foci. After surgery, P300 was still maximal over the vertex regardless of the side of resection but, in those patients who underwent STL, SAH and SH, its amplitude decreased only at the restricted area corresponding to the surgical intervention. In the MEG study, M400 was evoked at the anterior, middle and posterior lateral channels on each hemisphere. ECDs for M400 were detected in the mesial temporal area, the superior temporal area and the inferior parietal area on both hemispheres in the normal subjects as well as in the patients before surgery. After surgery, M400 disappeared only at the anterior lateral channels on the resected side in the patients who underwent STL, SAH and SH, but at other channels it did not differ from that observed before surgery. Only the ECD for M400 in the mesial temporal area on the resected side was lost. These findings suggest that the auditory processing is carried out by the combination of multiple structures including the mesial temporal area, the superior temporal area and the inferior parietal area on both hemispheres, and that the generator in the mesial temporal area contributes to the scalp-recorded P300 only at the corresponding anterior temporal region, but does not influence its predominant distribution over the scalp.
The scalp distributions of P300 in patients before surgery were consistent with those observed in the normal subjects studied as controls, and also with the previously reported data which were studied by employing both counting the number of target stimuli and button press as the response task (Vaughan and Ritter, 1970; Synder et al., 1980; Polich et al., 1991; Kanda et al., 1996). The results of the present topographic study in the patients after temporal lobectomy are consistent with some of the previous reports based on similar patients, which showed neither reduction in P300 amplitude nor altered scalp distribution of P300 that would be expected if P300 were generated solely from the mesial temporal area (Wood et al., 1982; Johnson and Fedio, 1986; Stapleton et al., 1987; Johnson, 1988). Studies in patients with amnesia showed that the mesial temporal formation did not contribute significantly to the scalp-recorded P300 (Onofrj et al., 1991, 1992; Polich and Squire, 1993; Honda et al., 1996). These studies, including the present one, indicate the importance of other main generator sources for the scalp-recorded P300. This assumption is supported by reports that lesions in the temporo-parietal and posterior association cortices resulted in decrements in the auditory and somatosensory P300 (Knight et al., 1989; Yamaguchi and Knight, 1991a, b, 1992). On the other hand, Meador et al. (1987), Wood et al. (1988) and Puce et al. (1989) showed, by invasive recordings, loss of limbic P3 in association with hippocampal sclerosis which was the most common histopathological finding in patients with temporal lobe epilepsy (Margerison and Corsellis, 1966; Babb and Brown, 1986, 1987; Puce et al., 1989; Dennis and Jaime, 1991), and that the unilaterally absent limbic P300 was a common finding in patients with intractable temporal lobe epilepsy. In these studies, however, the P3-like potentials directly recorded from the mesial temporal area were totally different from the scalp-recorded P300 in their waveforms. In the present study, even in five patients who were found to have hippocampal sclerosis, P300 showed normal distribution before surgery, and the focal attenuation on the operated side was seen only after surgery.
One possible explanation as to why the unilateral temporal lobectomy had little effect on the scalp-recorded P300 is related to the extent of propagation of the mesial temporal area activity over the scalp. The hippocampus is a laminated structure in which the synaptic current flow tends to summate rather than cancel, and therefore is suited anatomically for producing ERPs. Wood et al. (1984) and Halgren et al. (1986) raised the question of whether the scalp-recorded P300 was a volume-conducted electric field of the mesial temporal area activity or the product of other possible generators, but the extent to which the activity generated in the hippocampus and amygdala contributes to the field is unknown. Altafullah et al. (1986) found that a post-epileptic slow wave activity, which was generated in the same area of the mesial temporal area, propagated to the lateral cortical surface, and further reported that, in comparison with the attenuation of the slow wave, the surface P300 was about twice as large as would be predicted if the mesial temporal area were its sole generator. In the present study, there were no significant differences in the waveform, peak latency or amplitude of P300 measured at Pz in the temporal lobe epilepsy patients before and after surgery. These results indicate that the peak latency and amplitude of P300 measured at Pz are not affected by the unilateral partial resection of the temporal lobe, and rather suggest the involvement of multiple sites, including the mesial temporal area, in auditory cognitive processing. Therefore, P300 most likely reflects the summation of multiple, parallel, neural activities that form a network for analysing incoming information. Evidence in support of the existence of multiple generators was obtained from lesion studies (Halgren et al., 1986; Stapleton and Halgren, 1987; Johnson, 1988; Smith and Halgren, 1989; Nishitani et al., 1996), and from MEG study in normal subjects (Nishitani et al., 1998).
A theoretical difficulty exists in determining whether a lesion-induced change in the scalp field is the direct result of altering a generator source or not. Rugg et al. (1991a) recorded the auditory and visual oddball P300s on three different occasions from a patient with an infiltrating glioma in the left mesial temporal area, by using the balanced sterno-vertebral reference electrode. They reported that the amplitude of P300 at Pz was within normal limits and the scalp distribution of P300 over the central areas was symmetric, whereas the amplitude of P300 at the anterior electrodes on the affected side was altered on every occasion, suggesting that the scalp distribution and amplitude of P300 depend on the nature and the site of the lesions. Potter et al. (1993) studied a case of bilateral mesial temporal area lesions following viral encephalitis by using the same recording methods as Rugg et al. (1991a), and found that the auditory and visual oddball P300s did not show any abnormality at any electrodes. The brain CT and MRI in these studies showed certain damage in the mesial temporal area, but the lesions in these cases are obviously different from the temporal lobectomy, in which at least a part of possible generators of P300 is actually removed. Another plausible explanation for the discrepancy between the results of these studies and those of the present study may be the difference in the recording sites. Rugg et al. (1991a) and Potter et al. (1993) did not record from the lateral temporal electrodes. It is certainly possible that the lateral temporal electrodes might have shown an alteration of the P300 amplitude, as in the present study which showed the attenuation of the P300 amplitude only at the anterior temporal electrodes on the resected side.
Finally, patients reported by Onofrj et al. (1991, 1992) had extensive pathology in the bilateral mesial temporal area. Their scalp-recorded P300 showed a symmetric distribution, but the amplitude of P300 was attenuated at lateral temporal electrodes on each hemisphere. In their cases, however, recordings were made by using linked earlobes as the reference while the lesions extended beyond the mesial temporal area. These reference sites are very close to the lateral temporal scalp and thus might not be ideal for detecting potentials over the temporal region, even in cases without skull bone defect (Rugg, 1995).
In humans, there are methodological limitations in the study of the source number, localization and configuration of ERP. In order to clarify whether the hippocampus, amygdala and adjacent cortical areas play an essential role in generating the P300-like wave corresponding to the scalp-recorded P300, ERPs were studied in animal models; in monkey (Paller et al., 1982, 1988, 1992; Arthur et al., 1984; Neville and Foote, 1984; Glover et al., 1986), cat (Wilder et al., 1981; Buchwald and Squires, 1982; O'Connor and Starr, 1985), rabbit (Gabriel et al., 1983; Weisz et al., 1983) and rat (Ehlers et al., 1994). The late positive component of ERPs recorded in these animals was widely distributed, being maximal over the sensorimotor and parietal cortical areas, which conforms with the results in humans. P300 was recorded in the monkey after total bilateral amygdalohippocampectomy without causing significant change in the waveforms (Paller et al., 1988). Those ERPs were quite similar to the human P300 in terms of latency, polarity, topography, general waveforms and physiological characteristics. These results suggest that the late positive waves may reflect common neurophysiological processes in monkey and humans, and are consistent with the results in the present patients with temporal lobe epilepsy.
Two factors must be taken into account for interpreting the EEG data obtained after neurosurgical treatment of temporal lobe epilepsy. One is the extent of resection in the mesial temporal area and the other is the skull bone defect. Regarding the former, temporal lobectomy usually spares a substantial portion of the posterior hippocampus (Awad et al., 1989; Knight, 1990). In the present study, as the result of resection of the anterior hippocampus, the ipsilateral P300 was attenuated or disappeared, indicating that at least a part of the generator of P300 is located in the anterior mesial temporal area. However, there has been no agreement as to where exactly in the hippocampus the P300 response originates. Wood et al. (1984) found the polarity reversal of P300 in the posterior hippocampus. Rugg et al. (1991b) reported that the anterior mesial temporal area was not the principal locus for the generators of ERPs in the visual recognition task. On the other hand, according to the special criteria in which the dipoles were assumed in the cube of <300 mm3 (see Source modelling from MEG in Methods), the results identified by the present MEG study strongly suggest that the generating site in the mesial temporal area is located in its anterior portion. This finding is consistent with the result obtained by Halgren et al. (1986), who reported the consistent localization of the intracranial P300-like activity in the anterior mesial temporal area.
The second factor is how much the skull bone defect intervenes in the scalp-recorded EEG signals. The high impedance of the skull is a major influence on the current flow from brain to scalp (Nunez, 1981, 1995a, b). The skull defect may induce changes in the volume current flow, which can affect the scalp field distribution of ERP components. This explains the present finding that the distribution of N100 was asymmetric and was larger over the anterior quadrant on the operated hemisphere. In contrast, in spite of the predicted volume current flow, the amplitude of P300 was attenuated or even disappeared only over the limited area corresponding to the surgery, regardless of the type of operation, while the distribution of P300 over the vertex showed no significant difference before versus after surgery. There are previous studies suggesting that the local attenuation of P300 after STL was masked because of the increase in volume current flow caused by the skull defect (Stapleton et al., 1987; Johnson, 1988). Furthermore, Johnson (1988) recorded the P300 from the midline and parasagittal electrodes referenced to the linked earlobes. Therefore, their results might also have been affected by the recording electrode sites and the linked earlobe reference. In the present study, however, BNE was used as a reference in order to avoid the effect of current flow changes caused by skull defect on the reference electrode.
EEG does not have high spatial resolution, as described above, since it is easily affected by the different electric conductivity of head structures. The contribution of the mesial temporal area to the cognitive processing, which formerly was estimated from the results of topographical EEG studies, can be clarified by MEG, which, in contrast to EEG, can record brain activities non-invasively without being influenced by changes in the current flow caused by the surrounding head structures (Barth et al., 1986). The results of the present MEG studies clearly suggest that only the magnetic deflections over the limited area reflect the resection-induced change of the intracerebral generator source while those over other areas are not affected. Regarding the generator source in the mesial temporal area, the studies using the intracranial depth electrodes support the results obtained by the present MEG study. These studies showed that the intracranial ERPs varied in amplitude and polarity over small distances within and adjacent to the hippocampal formation and amygdala, suggesting that P300 was generated in or close to these structures (Halgren et al., 1980, 1995; Squires et al., 1983; Johnson and Fedio, 1986 Johnson and Fedio, 1987; McCarthy and Wood, 1987; Stapleton and Halgren, 1987; Stapleton et al., 1987; Halgren, 1988; Johnson, 1989a, b; McCarthy et al., 1989). Furthermore, the MEG system used in the present study employed the planar type neuromagnetometer which can detect the generating site just below the channels which show the maximum magnetic deflections. Therefore, the active areas identified by the present MEG study are considered to be highly reliable. The MEG results in the patients before surgery were exactly the same as those that we observed in the normal subjects (Nishitani et al., 1998). Further, they are partially supported by the previous MEG study (Okada et al., 1983; Gordon et al., 1987; Lewine et al., 1990; Rogers et al., 1991; Tesche et al., 1996). Judging from the present results studied by MEG, three dipoles detected in the mesial temporal area, the superior temporal area and the inferior parietal area on each hemisphere are considered to be active almost simultaneously in the auditory stimulus detection task. The results relating to M400 and its ECDs indicate that the auditory cognitive processing is carried out by the combination of multiple regions described above, and that the unilateral resection of the mesial temporal area does not affect this processing. Consequently, the results of ERFs in the present study support those of ERPs.
In two patients (Patients 4 and 9), P300 was easily recognized whereas M400 was not detected at any channel. It is conceivable that, in those cases, the dipole was directed radially with respect to the head surface at the latency of the M400 peak, because MEG is most sensitive to the tangentially oriented dipole and much less sensitive to other dipoles such as randomly or radially oriented ones (Cohen and Cuffin, 1983; Lopes da Silva et al., 1991, 1993). Further, the reason why the degree of difference in magnetic responses between the target and non-target stimuli varied among subjects is presumed to be that the channels showing the largest magnetic signals to the target stimuli in each subset channel were not always the same before and after surgery or among subjects, and that the cognitive information processes are not entirely the same among subjects. The latter reason is supported by the fact that the peak latency of M400 was variable among different areas and among subjects.
|
Acknowledgments |
|---|
This work was supported by Grants-in-Aid for Scientific Research (A) 09308031, (A) 08558083, (C) 09670655 and (C) 10670583, and on Priority Areas 08279106, and for International Scientific Research 07044258 from the Japan Ministry of Education, Science, Sports and Culture, Research for the Future Program JSPS-RFTF 97L00201 from the Japan Society for the Promotion of Science, and General Research Grant for Aging and Health from the Japan Ministry of Health and Welfare.
|
References |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Ahonen AI, Hämäläinen MS, Kajola MJ, Knuutila JET, Laine PP, Lounasmaa OV, et al. 122-channel SQUID instrument for investigating the magnetic signals from the human brain. Physica Scripta 1993; T49: 198–205.[Web of Science]
Altafullah I, Halgren E, Stapleton JM, Crandall PH. Interictal spike-wave complexes in the human medial temporal lobe: typical topography and comparisons with cognitive potentials. Electroencephalogr Clin Neurophysiol 1986; 63: 503–16.[Web of Science][Medline]
Arthur DL, Starr A. Task-relevant late positive component of the auditory event-related potential in monkeys resembles P300 in humans. Science 1984; 233: 186–8.
Awad IA, Katz A, Hahn JF, Kong AK, Ahl J, Lüders H. Extent of resection in temporal lobectomy for epilepsy. I. Interobserver analysis and correlation with seizure outcome. Epilepsia 1989; 30: 756–62.[Web of Science][Medline]
Babb TL, Brown WJ. Neuronal, dendritic, and vascular profile of human temporal lobe epilepsy correlated with cellular physiology in vivo. Adv Neurol 1986; 44: 949–66.[Medline]
Babb TL, Brown WJ. Pathological findings in epilepsy. In: Engel J Jr, editor. Surgical treatment of the epilepsies. New York: Raven Press; 1987. p. 520–4.
Barrett G, Neshige R, Shibasaki H. Human auditory and somatosensory event-related potentials: effects of response condition and age. Electroencephalogr Clin Neurophysiol 1987; 66: 409–19.[Web of Science][Medline]
Barth DS, Sutherling W, Broffman J, Beatty J. Magnetic localization of a dipolar current source implanted in a sphere and a human cranium. Electroencephalogr Clin Neurophysiol 1986; 63: 260–73.[Web of Science][Medline]
Buchwald JS, Squires NS. Endogenous auditory potentials in the cat: a P300 model. In: Woody CD, editor. Conditioning: representation of involved neural function. New York: Plenum Press; 1982. p. 503–15.
Cohen D, Cuffin BN. Demonstration of useful differences between magnetoencephalogram and electroencephalogram. Electroencephalogr Clin Neurophysiol 1983; 56: 38–51.[Web of Science][Medline]
Commission on Classification and Terminology of the International League Against Epilepsy. Proposal for revised classification of epilepsies and epileptic syndromes. Epilepsia 1989; 30: 389–99.[Web of Science][Medline]
Courchesne E, Hillyard SE, Courchesne RY. P3 waves to the discrimination of targets in homogeneous and heterogeneous stimulus sequences. Psychophysiology 1977; 14: 590–7.[Web of Science][Medline]
Dennis DS, Jaime I. Temporal lobectomy. In: Lüders H, editor. Epilepsy surgery. New York: Raven Press; 1991. p. 533–45.
Donchin E, Coles MG. On the conceptual foundations of cognitive psychophysiology. Behav Brain Sci 1988; 11: 408–27.[Web of Science]
Donchin E, Ritter W, McCallum WC. Cognitive psychophysiology: the endogenous components of the ERP. In: Callaway E, Tueting P, Koslow SH, editors. Event-related brain potentials in man. New York: Academic Press; 1978. p. 349–411.
Ehlers CL, Kaneko WM, Robledo P, Lopez AL. Long-latency event-related potentials in rats: effects of task and stimulus parameters. Neuroscience 1994; 62: 759–69.[Web of Science][Medline]
Fabiani M, Gratton G, Karis D, Donchin E. The definition, identification and reliability of measurement of the P300 component of the event-related brain potential. In: Ackles PK, Jennings JR, Coles MG, editors. Advances in Psychophysiology, Vol. II. Greenwich (CT): JAI Press; 1987. p. 1–78.
Gabriel M, Sparenborg SP, Donchin E. Macropotentials recorded from the cingulate cortex and anterior thalamus during the `oddball' paradigm used to elicit P300 in humans [abstract]. Soc Neurosci Abstr 1983; 9: 1200.
Glover AA, Onofrj MC, Ghilardi MF, Bodis-Wollner I. P300-like potentials in the normal monkey using classical conditioning and an auditory `oddball' paradigm. Electroencephalogr Clin Neurophysiol 1986; 65: 231–5.[Web of Science][Medline]
Gordon E, Sloggett G, Harvey I, Kraiuhin C, Rennie C, Yiannikas C, Meares R. Magnetoencephalography: locating the source of P300 via magnetic field recording. Clin Exp Neurol 1987; 23: 101–10.[Medline]
Halgren E. The P3: a view from the brain. Behav Brain Sci 1988; 11: 357–74.[Web of Science]
Halgren E, Squires NK, Wilson CL, Rohrbaugh JW, Babb TL, Crandall PH. Endogenous potentials generated in the human hippocampal formation and amygdala by infrequent events. Science 1980; 210: 803–5.
Halgren E, Stapleton JM, Smith MS, Altafullah I. Generators of the human scalp P3(s). In: Cracco RQ, Bodis-Wolner I, editors. Evoked potentials, III. Boston (MA): Butterworths; 1986. p. 269–84.
Halgren E, Baudena P, Clarke JM, Heit G, Liegeois C, Chauvel P, et al. Intracerebral potentials to rare target and distractor auditory and visual stimuli. I. Superior temporal plane and parietal lobe. Electroencephalogr Clin Neurophysiol 1995; 94: 191–220.[Web of Science][Medline]
Hämäläinen M, Hari R, Ilmoniemi R, Knuutila J, Lounasmaa OV. Magnetoencephalography—theory, instrumentation, and applica- tions to noninvasive studies of the working human brain. Rev Mod Phys 1993; 65: 413–97.[Web of Science]
Hari R, Karhu J, Hämäläinen M, Knuutila J, Salonen O, Sams M, et al. Functional organization of the human first and second somatosensory cortices: a neuromagnetic study. Eur J Neurosci 1993; 5: 724–34.[Web of Science][Medline]
Honda M, Suwazono S, Nagamine T, Yonekura Y, Shibasaki H. P300 abnormalities in patients with selective impairment of recent memory. J Neurol Sci 1996; 139: 95–105.
Ilmoniemi R. Estimates of neuronal current distributions. Acta Otolaryngol (Stockh) 1991; Suppl 491: 80–7.
Johnson R Jr. Scalp-recorded P300 activity in patients following unilateral temporal lobectomy. Brain 1988; 111: 1517–29.
Johnson R Jr. Auditory and visual P300s in temporal lobectomy patients: evidence for modality-dependent generators. Psychophysiology 1989a; 26: 633–50.[Web of Science][Medline]
Johnson R Jr. Developmental evidence for modality-dependent P300 generators: a normative study. Psychophysiology 1989b; 26: 651–67.[Web of Science][Medline]
Johnson R Jr, Fedio P. P300 activity following unilateral temporal lobectomy: a preliminary report. Electroencephalogr Clin Neurophysiol 1986; Suppl 38: 552–4.
Johnson R Jr, Fedio P. Task-related changes in P300 scalp distribution in temporal lobectomy patients. Electroencephalogr Clin Neurophysiol 1987; Suppl 40: 699–704.
Kanda M, Fujiwara N, Xu X, Shindo K, Nagamine T, Ikeda A, et al. Pain-related and cognitive components of somatosensory evoked potentials following CO2 laser stimulation in man. Electroencephalogr Clin Neurophysiol 1996; 100: 105–14.[Medline]
Kaukoranta E, Hari R, Hämäläinen M, Huttunen J. Cerebral magnetic fields evoked by peroneal nerve stimulation. Somatosens Res 1986; 3: 309–21.[Web of Science][Medline]
Knight RT. Neural mechanism of event-related potentials: evidence from human lesion studies. In: Rohrbaugh JW, Parasuraman R, Johnson R, editors. Event-related brain potentials: issues and interdisciplinary vantages. New York: Oxford University Press; 1990. p. 3–18.
Knight RT, Scabini D, Woods DL, Clayworth CC. Contributions of temporal–parietal junction to the human auditory P3. Brain Res 1989; 502: 109–16.[Web of Science][Medline]
Levänen S, Ahonen A, Hari R, McEvoy L, Sama M. Deviant auditory stimuli activate human left and right auditory cortex differently. Cereb Cortex 1996; 6: 288–96.
Lewine JD, Roeder SB, Oakey MT, Arthur DL, Aine CJ, George JS, et al. Localization of the generators of the magnetic P3. In: Williamson S, Hoke M, Stroink G, Kotani M, editors. Advances in biomagnetism. New York: Plenum Press; 1990. p. 229–38.
Lopes da Silva FH, Rotterdam AV. Biophysical aspects of EEG and magnetoencephalogram generation. In: Niedermeyer E, Lopes da Silva FH, editors. Electroencephalography: basic principles, clinical applications and related fields. Baltimore (MD): Williams & Wilkins; 1993. p. 78–91.
Lopes da Silva FH, Wieringa HJ, Peters MJ. Source localization of EEG versus MEG: empirical comparison using visually evoked responses and theoretical considerations. Brain Topogr 1991; 4: 133–42.[Medline]
Margerison JH, Corsellis JA. Epilepsy and the temporal lobe. Brain 1966; 89: 499–530.
McCarthy G, Wood CC. Scalp distributions of event-related potentials: an ambiguity associated with analysis of variance models. Electroencephalogr Clin Neurophysiol 1985; 62: 203–8.[Web of Science][Medline]
McCarthy G, Wood CC. Intracranial recordings of endogenous ERPs in humans. Electroencephalogr Clin Neurophysiol 1987; Suppl 39: 331–7.
McCarthy G, Wood CC, Williamson PD, Spencer DD. Task-dependent field potentials in human hippocampal formation. J Neurosci 1989; 9: 4253–68.[Abstract]
Meador KJ, Loring DW, King DW, Gallagher BB, Gould MJ, Flanigin HF, et al. Limbic evoked potentials predict site of epileptic focus. Neurology 1987; 37: 494–7.
Mosher JC, Lewis PS, Leahy RM. Multiple dipole modeling and localization from spatio-temporal MEG data. IEEE Trans Biomed Eng 1992; 39: 541–57.[Web of Science][Medline]
Neshige R, Lüders H. Recording of event-related potentials (P300) from human cortex. J Clin Neurophysiol 1992; 9: 294–8.[Web of Science][Medline]
Neville HJ, Foote SL. Auditory event-related potentials in the squirrel monkey: parallels to human late wave responses. Brain Res 1984; 298: 107–16.[Web of Science][Medline]
Nishitani N, Ikeda A, Honda M, Nagamine T, Shibasaki H. Auditory event-related potentials in patients with intractable temporal lobe epilepsy before and after surgery. In: Kato N, editor. The hippocampus: functions and clinical relevance. Amsterdam: Elsevier; 1996. p. 415–8.
Nishitani N, Nagamine T, Fujiwara N, Yazawa S, Shibasaki H. Cortical–hippocampal auditory processing identified by magnetoencephalography. J Cogn Neurosci 1998; 10: 231–47.[Web of Science][Medline]
Nunez PL. Electric fields of the brain. The neurophysics of EEG. New York: Oxford University Press; 1981.
Nunez, PL. Quantitative states of neocortex. In: Nunez PL, editor. Neocortical dynamics and human EEG rhythms. New York: Oxford University Press; 1995a. p. 3–67.
Nunez, PL. Toward a physics of neocortex. In: Nunez PL, editor. Neocortical dynamics and human EEG rhythms. New York: Oxford University Press; 1995b. p. 68–132.
O'Connor TA, Starr A. Intracranial potentials correlated with an event-related potential, P300, in the cat. Brain Res 1985; 339: 27–38.[Web of Science][Medline]
O'Donnell BF, Cohen RA, Hokama H, Cuffin BN, Lippa C, Shenton ME, et al. Electrical source analysis of auditory ERPs in medial temporal lobe amnestic syndrome. Electroencephalogr Clin Neurophysiol 1993; 87: 394–402.[Web of Science][Medline]
Okada YC, Kaufman L, Williamson SJ. The hippocampal formation as a source of the slow endogenous potentials. Electroencephalogr Clin Neurophysiol 1983; 55: 417–26.[Web of Science][Medline]
Onofrj M, Gambi D, Fulgente T, Bazzano S, Colamartino P. Persistence of a P3 component in severe amnestic syndrome. Electroencephalogr Clin Neurophysiol 1991; 78: 480–84.[Web of Science][Medline]
Onofrj M, Fulgente T, Nobilio D, Malatesta G, Bazzano S, Colamartino P, et al. P3 recordings in patients with bilateral temporal lobe lesions. Neurology 1992; 42: 1762–7.
Paller KA, Zola-Morgan S, Squire LR, Hillyard SA. Late positive event-related potentials in cynomolgus monkeys (Macaca fasicularis) [abstract]. Soc Neurosci Abstr 1982; 8: 975.
Paller KA, Zola-Morgan S, Squire LR, Hillyard SA. P3-like brain waves in normal monkeys and in monkey with medial temporal lesions. Behav Neurosci 1988; 102: 714–25.[Web of Science][Medline]
Paller KA, McCarthy G, Roessler E, Allison T, Wood CC. Potentials evoked in human and monkey medial temporal lobe during auditory and visual oddball paradigms. Electroencephalogr Clin Neurophysiol 1992; 84: 269–79.[Web of Science][Medline]
Picton TW, Hillyard SA. Endogenous event-related potentials. In: Picton TW, editor. Handbook of electroencephalography and clinical neurophysiology, Vol. 3. Human event-related potentials. Amsterdam: Elsevier; 1988. p. 361–426.
Polich J, Squire LR. P300 from amnesic patients with bilateral hippocampal lesions. Electroencephalogr Clin Neurophysiol 1993; 86: 408–17.[Web of Science][Medline]
Polich J, Brock T, Geisler MW. P300 from auditory and somatosensory stimuli: probability and inter-stimulus interval. Int J Psychophysiol 1991; 11: 219–23.[Web of Science][Medline]
Potter DD, Pickles CD, Roberts RC, Rugg MD, Paller KA, Mayes AR. Visual and auditory P300 in a case of bilateral destruction of the medial temporal lobes following viral encephalitis. In: Heinze H, Munte TF, Mangun GR, editors. New developments in event-related potentials. Boston: Birkhäuser; 1993. p. 303–11.
Powell GE, Polkey CE, Canavan AG. Lateralization of memory functions in epileptic patients by use of the sodium amytal (Wada) technique. J Neurol Neurosurg Psychiatry 1987; 50: 665–72.
Puce A, Kalnins RM, Berkovic SF, Donnan GA, Bladin PF. Limbic P3 potentials, seizure localization, and surgical pathology in temporal lobe epilepsy. Ann Neurol 1989; 26: 377–85.[Web of Science][Medline]
Rogers RL, Baumann SB, Papanicolaou AC, Bourbon TW, Alagarsamy S, Eisenberg HM. Localization of the P3 sources using magnetoencephalography and magnetic resonance imaging. Electroencephalogr Clin Neurophysiol 1991; 79: 308–21.[Web of Science][Medline]
Rugg MD. Cognitive event-related potentials: intracranial and lesion studies. In: Boller F, Grafman J, editors. Handbook of neuropsychology, Vol. 10. Amsterdam: Elsevier; 1995. p. 165–85.
Rugg MD, Pickles CD, Potter DD, Roberts RC. Normal P300 following extensive damage to the left medial temporal lobe. J Neurol Neurosurg Psychiatry 1991a; 54: 217–22.
Rugg MD, Roberts RC, Potter DD, Pickles CD, Nagy M. Event-related potentials related to recognition memory. Brain 1991b; 114: 2313–32.
Scherg M. Functional imaging and localization of electromagnetic brain activity. Brain Topogr 1992; 5: 103–11.[Medline]
Scherg M, Von Cramon D. Evoked dipole source potentials of the human auditory cortex. Electroencephalogr Clin Neurophysiol 1986; 65: 344–60.[Web of Science][Medline]
Smith ME, Halgren E. Dissociation of recognition memory components following temporal lobe lesions. J Exp Psychol: Learn, Mem Cogn 1989; 15: 50–60.[Web of Science][Medline]
Smith ME, Stapleton JM, Halgren E. Human medial temporal lobe potentials evoked in memory and language tasks. Electroencephalogr Clin Neurophysiol 1986; 63: 145–59.[Web of Science][Medline]
Squires NK, Squires KC, Hillyard SA. Two varieties of long-latency positive waves evoked by unpredictable auditory stimuli in man. Electroencephalogr Clin Neurophysiol 1975; 38: 387–401.[Web of Science][Medline]
Squires KC, Donchin E, Herning RI, McCarthy G. On the influence of task relevance and stimulus probability on event-related-potential components. Electroencephalogr Clin Neurophysiol 1977; 42: 1–14.[Web of Science][Medline]
Squires NK, Halgren E, Wilson C, Crandall P. Human endogenous limbic potentials: cross-modality and depth/surface comparisons in epileptic subjects. In: Gaillard AW, Ritter W, editors. Tutorials in event related potential research: endogenous components. Amsterdam: North-Holland; 1983. p. 217–32.
Stapleton JM, Halgren E. Endogenous potentials evoked in simple cognitive tasks: depth components and task correlates. Electroencephalogr Clin Neurophysiol 1987; 67: 44–52.[Web of Science][Medline]
Stapleton JM, Halgren E, Moreno KA. Endogenous potentials after temporal lobectomy. Neuropsychologia 1987; 25: 549–57.[Web of Science][Medline]
Stephenson WA, Gibbs FA. A balanced non-cephalic reference electrode. Electroencephalogr Clin Neurophysiol 1951; 3: 237–40.
Suwazono S, Shibasaki H, Nishida S, Nakamura M, Honda M, Nagamine T, et al. Automatic detection of P300 in single sweep records of auditory event-related potential. J Clin Neurophysiol 1994; 11: 448–60.[Web of Science][Medline]
Synder E, Hillyard SA, Galambos R. Similarities and differences among the P3 waves to detected signals in three modalities. Psychophysiology 1980; 17: 112–22.[Web of Science][Medline]
Tesche CD, Karhu J, Tissari SO. Non-invasive detection of neuronal population activity in human hippocampus. Brain Res Cogn Brain Res 1996; 4: 39–47.[Medline]
Vaughan HG Jr, Ritter W. The sources of auditory evoked responses recorded from the human scalp. Electroencephalogr Clin Neurophysiol 1970; 28: 360–7.[Web of Science][Medline]
Wada J, Rasmussen T. Intracarotid injection of sodium amytal for the lateralization of cerebral speech dominance: experimental and clinical observations. J Neurosurg 1960; 17: 266–82.[Web of Science]
Weisz DJ, McCarthy G, Wood CC, Thompson DT. Event-related potentials reflect stimulus significance during discriminative NM conditioning in the rabbit [abstract]. Soc Neurosci Abstr 1983; 9: 642.
Wilder MB, Farley GR, Starr A. Endogenous late positive component of the evoked potential in cats corresponding to P300 in humans. Science 1981; 211: 605–7.
Wood CC, McCarthy G, Allison T, Goff WR, Williamson PD, Spencer DD. Endogenous event-related potentials following temporal lobe excisions in humans [abstract]. Soc Neurosci Abstr 1982; 8: 976.
Wood CC, McCarthy G, Squires NK, Vaughan HG, Woods DL, McCallum WC. Anatomical and physiological substrates of event-related potentials. Two case studies. Ann NY Acad Sci 1984; 425: 681–721.[Web of Science][Medline]
Wood CC, McCarthy G, Kim JH, Spencer DD, Williamson PD. Abnormalities in temporal lobe event-related potentials predict hippocampal cell loss in temporal lobe epilepsy [abstract]. Soc Neurosci Abstr 1988; 14: 5.
Yamaguchi S, Knight RT. P300 generation by novel somatosensory stimuli. Electroencephalogr Clin Neurophysiol 1991a; 78: 50–5.[Web of Science][Medline]
Yamaguchi S, Knight RT. Anterior and posterior association cortex contributions to the somatosensory P300. J Neurosci 1991b; 11: 2039–54.[Abstract]
Yamaguchi S, Knight RT. Effects of temporal–parietal lesions on the somatosensory P3 to lower limb stimulation. Electroencephalogr Clin Neurophysiol 1992; 84: 139–48.[Web of Science][Medline]
Received March 19, 1998. Revised September 30, 1998. Second revision on November 23, 1998. Accepted November 30, 1998.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  D. Papo, A. Douiri, F. Bouchet, J.-C. Bourzeix, J.-P. Caverni, and P.-M. Baudonniere Time-Frequency Intracranial Source Localization of Feedback-Related EEG Activity in Hypothesis Testing Cereb Cortex, June 1, 2007; 17(6): 1314 - 1322. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. D. Tesche and J. Karhu Theta oscillations index human hippocampal activation during a working memory task PNAS, January 18, 2000; 97(2): 919 - 924. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||