PMCCPMCCPMCC

Search tips
Search criteria 

Advanced

 
Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
 
Ann Neurol. Author manuscript; available in PMC 2012 March 12.
Published in final edited form as:
PMCID: PMC3299311
NIHMSID: NIHMS135473

3D surface maps link local atrophy & Fast Ripples in human epileptic hippocampus

Abstract

Objective

There is compelling evidence that pathological high frequency oscillations (HFOs) called Fast Ripples (FR, 150–500 Hz) reflect abnormal synchronous neuronal discharges in areas responsible for seizure genesis in patients with mesial temporal lobe epilepsy (MTLE). It is hypothesized that morphological changes associated with hippocampal atrophy (HA) contribute to the generation of FR, yet there is limited evidence that hippocampal FR-generating sites correspond with local areas of atrophy.

Methods

Interictal HFOs were recorded from hippocampal microelectrodes in ten patients with MTLE. Rates of FR and Ripple discharge from each microelectrode were evaluated in relation to local measures of HA obtained using 3D MRI hippocampal modeling.

Results

Rates of FR discharge were three times higher in areas of significant local HA compared to rates in non-atrophic areas. Furthermore, FR occurrence correlated directly with the severity of damage in these local atrophic regions. In contrast, we found no difference in rates of Ripple discharge between local atrophic and non-atrophic areas.

Interpretation

The proximity between local HA and microelectrode-recorded FR suggest morphological changes such as neuron loss and synaptic reorganization may contribute to the generation of FR. Pathological HFOs, such as FR, may provide a reliable surrogate marker of abnormal neuronal excitability in hippocampal areas responsible for the generation of spontaneous seizures in patients with MTLE. Based on these data, it is possible that MRI-based measures of local HA could identify FR-generating regions, and thus provide a non-invasive means to localize epileptogenic regions in hippocampus.

Introduction

High frequency oscillations (HFOs) greater than 80 Hz have attracted much attention for their potential role in information processing, and more recently, in neurological disease in the mammalian brain. Early studies on Ripple oscillations (100–200 Hz) showed that these HFOs occurred bilaterally in hippocampus and parahippocampal structures of naïve rodents.13 Later studies described Ripples in non-human primates and patients with epilepsy that occurred as intermittent, brief bursts (10–100 msec) in mesial temporal lobe structures, but were slightly lower in spectral frequency compared to rats (i.e. 80–150 Hz).46 In non-primates, and possibly humans, Ripples may be associated with information transfer between hippocampus and extra-hippocampal structures,711 but the occurrence and function of Ripples in epileptogenic regions of patients with epilepsy is unclear.1217

Fast Ripples (FR) are another type of HFO that are typically higher in spectral frequency than Ripples, and may contain frequencies as high as 600Hz. FR are strongly associated with brain areas of epileptic seizure onset,4, 5, 13, 15, 18, 19 and sometimes occur immediately before or during the onset of mesial temporal lobe seizures.12, 20 FR are believed to reflect neuronal disturbances responsible for epileptogenicity.21 Studies in patients with mesial temporal lobe epilepsy (MTLE) and rat models of human MTLE suggest that hippocampal atrophy (HA) may be the underlying anatomical disturbance contributing to FR generation. There is evidence that reduced hippocampal volumes and lower neuron densities correlate with higher rates of FR and lower rates of Ripple occurrence.15, 16, 22 It is also known, however, that cell loss is often greater within specific hippocampal subfields,2325 and may not be distributed evenly along the anterior-posterior axis of the epileptic hippocampus.2628 Furthermore, FR may arise from local areas that are distributed unevenly throughout the hippocampus in epileptic rats and patients with MTLE.29, 30 It is possible, then, that areas of HA may not be in close proximity to areas supporting FR generation, which would argue against our hypothesis that morphological changes associated with HA contribute to FR generation.16

In order to evaluate the relative proximity or distance between areas of HA and FR-generating areas, we used advanced MRI analysis techniques that quantify the distribution of atrophy in hippocampal three-dimensional (3D) reconstructed models.27, 28, 3134 Moreover, in the present study using surgical patients with medically intractable MTLE, we have extended these MRI techniques by registering the position of hippocampal intracranial microelectrodes and corresponding rates of FR and Ripple occurrence, on these 3D maps of HA in order to evaluate whether local HA are associated with areas supporting the generation of FR or Ripples or both.

Subjects and Methods

Patients and Electrophysiological Recordings

Subjects were patients with medically intractable seizures of probable temporal lobe origin, who were candidates for epilepsy surgery, but required intracranial depth electrode evaluation to localize brain areas where seizures began because results from non-invasive studies suggesting focal onsets were inconclusive. In each patient, flexible polyurethane depth electrodes (AdTech Medical Instruments, Racine, WI) were implanted bilaterally (median number of electrodes per patient: 5.0 right, 4.5 left hemisphere) in temporal and frontal lobe areas, orthogonal to the lateral skull surface, in order to identify brain areas generating spontaneous seizure activity (Fig. 1A).35, 36 Each depth electrode was 1.3 mm in diameter and consisted of 7 contacts (1.5 mm length) with inter-contact center-to-center spacing of 6 mm, except between the two most distal contacts where spacing was 3 mm. Brain areas of electrographic seizure onset were identified by attending neurologists at the UCLA Seizure Disorders Center following review of multiple spontaneous, independent seizure recordings. Because the objective of this retrospective study was to evaluate HFO activity in relation to areas of hippocampal atrophy and epileptogenicity, patients included in this study had unilateral mesial temporal lobe seizure onsets, evidence of hippocampal HFOs in continuous microelectrode recordings, and a complete series of post-implant MRI scans for microelectrode localization. All patients gave their informed consent to participate in this study, which was approved by the Medical Institutional Review Board of the UCLA Office for Protection of Research Subjects.

Figure 1
Interictal high frequency oscillations recorded from microelectrodes in the human hippocampus. (A) Coronal post-implant MRI (left) showing depth electrode positioned in left temporal lobe orthogonal to lateral surface of temporal bone. Note that the area ...

Inserted through the lumen of each depth electrode was a bundle of 9 platinum-iridium microwires (40 µm diameter; impedance 100–300 kOhm at 1 kHz) that extended 3–5 mm beyond the tip (Fig. 1A). The 9th microwire in each bundle was uninsulated (impedance 1–3 kOhm at 1kHz), and used as a reference. For each patient, wide bandwidth (0.1 Hz to 5 kHz) depth EEG was recorded from 16 microelectrodes simultaneously (10 kHz sampling; 12-bit precision, R.C. Electronics, Santa Barbara CA) during overnight polysomnographic sleep studies as part of a previous study.15 Because the highest probability for HFO discharge in non-primates3, 37, 38 and humans19, 39 occurs during episodes of slow wave sleep, unfiltered and bandpass filtered EEG (80–500 Hz; finite impulse response filter, 513 components) from each hippocampal microelectrode (n=72) was reviewed for HFO activity (Figs. 1B, 1C) during a ten minute epoch of slow wave sleep at 500 ms/page in a computer display window (Run Technologies, Co., Mission Viejo, CA). For microelectrodes with evidence of HFOs (18 out of 72 microelectrodes), the entire continuous recording (median duration of recording 4.4 hrs, range 3.0 to 6.1 hrs) was processed using a semi-automated computer algorithm in order to detect and quantify HFOs.15 Wide bandwidth EEG was bandpass filtered between 80 and 500 Hz, and the root mean square (RMS) amplitude using a sliding 3 ms window computed from the filtered signal. Criteria for HFO, which included visual confirmation, were as follows: consecutive RMS amplitude values greater than 5 standard deviations (SD) above the grand mean RMS amplitude, longer than 6 ms in duration, and more than six peaks exceeding 3 SD above the overall mean amplitude of the rectified bandpass filtered signal. Power spectral analysis (1024 point Fast Fourier transform with zero padding, and Hamming window) was used to identify spectral frequency corresponding to maximum power for each HFO (Figs. 1B, 1C). Using spectral frequency criteria from previous quantitative studies that separated human Ripples and FR,15, 19 in the present study, HFOs with maximum power corresponding with a peak spectral frequency between 80 and 150 Hz were labeled Ripples, while HFOs with a peak spectral frequency between 151 and 500 Hz were labeled FR. For each microelectrode that captured HFOs, a mean rate of Ripple and FR discharge per minute was computed as the total number of Ripples or FR divided by length of recording time.

Neuroimaging and Hippocampal 3D Reconstruction

Whole brain MRI scans from patients (n=10; 4 female; overall mean age: 35.5 +/− 10.2 years) were acquired in the axial plane using a 1.5 T Siemens Sonata full body scanner with head coil. 3D T1-weighted images were acquired using a spoiled gradient recalled sequence (256 × 256 × 124 matrix; 1mm isotropic voxels; field of view, 28 cm; echo time, 9 ms; repetition time, 40 ms). Control subjects without history of neurological disease (n=19; 6 female, mean age: 29.9 +/− 3.9 years) were scanned using equivalent scan parameters on a different 1.5T MR scanner.

3D MRI scans from each subject were linearly registered to the ICBM53 (International Consortium for Brain Mapping) average brain template in a semi-automated fashion,40 and subsequently transformed into standard space, reoriented, resampled, and resliced in the coronal plane. The dentate gyrus, hippocampus proper, presubiculum, and subiculum were identified visually on each coronal slice using a standard neuroanatomic atlas,41 and were included in the hippocampal tracings. The boundaries of each hippocampus were delineated manually by a single experimenter blind to hemisphere of seizure onset (Figure 2A), according to criteria adapted from the Insausti & Pitkanen volumetric analysis and LONI hippocampal segmentation protocols.42 A 3D parametric mesh model of each traced hippocampus (n=58) was created using established hippocampal modeling methods.27, 34, 43, 44 Each model consisted of 30,000 points, distributed across the hippocampal surface in a spatially normalized manner.45 The distance from each of the surface points to the medial curve, which runs through the center of the hippocampus along its longitudinal axis, is known as radial distance, and is a measure of hippocampal “thickness” (Figures 2B & 2C). Inter-rater reliability, based on an analysis of 10 whole hippocampal volumes traced by two investigators, was strong between investigators using the tracing protocol with small differences in whole volume (single measurement, model 2 intraclass correlation = 0.94, F=2.47).4446

Figure 2
3D Hippocampal Surface Modeling. (A) The hippocampus is traced manually in consecutive coronal MRI slices. From the hippocampal tracings, a 3D hippocampal model (B) is constructed. The medial curve (represented in green) threads through the hippocampus ...

To locate each microelectrode in the 3D hippocampus, post-implant MRI was linearly registered to the same pre-implant MRI that was used to construct the 3D hippocampus. The tip of the microelectrode viewed on coronal post-implant scans was identified (e.g. Figure 1A), and the x, y, and z coordinates corresponding to the tip were registered to the 3D hippocampus. Hippocampal areas within a 5 mm radius of each microelectrode tip (Figure 2D), and in a separate analysis 3 mm radius, were used to evaluate hippocampal thickness in relation to HFO activity. These distances were consistent with results from previous patient studies that suggested HFOs could be recorded from distances up to 5 mm,13, 20, 30 and in the present study, were comparable with the overall mean radial distance of the patient hippocampus.

Data Analysis

Patient hippocampi were separated into ‘ipsilateral’ and ‘contralateral’ groups based on each patient’s hemisphere of ictal onset, which were compared to hippocampi from the corresponding hemisphere in the entire control group that consisted of similar proportion of male to female subjects, and similar mean age. ANOVA was used to statistically evaluate differences in hippocampal thickness between patients and controls at each homologous surface point in relation to side of ictal onset. Probability (P) values corresponding to the ANOVA F ratio at each surface point were color-coded and mapped onto the surface of 3D hippocampus to depict the distribution of statistically significant atrophy in the patient group.27, 33, 45 Areas on the 3D P maps corresponding to P < 0.05 were labeled ‘atrophic’, while areas associated with P ≥ 0.05 were labeled ‘non-atrophic’. Permutation testing was used to correct for multiple comparisons,47 and determine the likelihood that the observed proportion of suprathreshold P map statistics (P < 0.01) could occur by chance.45, 48 The number of permutations N was chosen to be 1 × 106 to control for the standard error SEp of omnibus probability p, which follows a binomial distribution B(N, p) with known standard error.49 The margin of error (95% confidence interval) for p is approximately 5% when N = 8,000. To further improve upon this, we ran 100,000 permutations, and chose a significance level of 0.01. Microelectrode data from all patients were combined and assigned to one ipsilateral and one contralateral hippocampus. In hippocampal areas where microelectrode data overlapped from multiple patients, a distance-weighted average of rates of FR or Ripple discharge, with respect to the surface, was computed. Student t-tests were used to compare whole hippocampal volumes within and between patient and control groups, and rates of Ripple and FR occurrence in local atrophic and non-atrophic areas. In the latter analysis, the t-statistic was adjusted using within and between group correlation coefficients to compensate for the potential effects of non-independent samples in the parametric analysis.50, 51 Correlation analysis was used to evaluate hippocampal volume reductions and magnitude of local atrophy, computed as a ratio of patient to control mean radial distance, in relation to rates of Ripple and FR discharge.

Results

Ten patients with unilateral mesial temporal lobe seizure onsets, and 19 age and gender matched subjects without history of neurological disorder, participated in this study. Consistent with the area of seizure onset and incidence of HS in these patients (Table 1), quantitative MRI analysis revealed that patients had significantly smaller ipsilateral hippocampal volumes compared to control subjects (patients’ mean volume ± SE: 2607 ± 220 mm3, controls’: 3621 ± 86 mm3, mean volume reduction: 29 ± 6%, p < 0.001), but contralateral volumes were not statistically different than controls (patients: 3531 ± 167 mm3, controls: 3698 ± 90 mm3, mean volume reduction: 4 ± 6%, p < 0.39). Within the patient group, ipsilateral hippocampi were 26 ± 6% smaller than hippocampi contralateral to onset (p < 0.004). No such asymmetry was observed within the control group (2 ± 2%, p < 0.54).

Table 1
Patient clinical variables

In patients, the mean rate of FR occurrence recorded from microelectrodes ipsilateral to seizure onset was twice as high as rates recorded contralaterally (0.42 ± 0.19 per min vs. 0.21 ± 0.15 per min). The mean rate of Ripple occurrence was three times higher in contralateral compared to ipsilateral hippocampi (0.12 ± 0.07 per min vs. 0.04 ± 0.02 per min). Analysis of hippocampal volumes in relation to rates of HFO discharge revealed that greater hippocampal volume reductions correlated with higher FR rates (r = 0.53, p < 0.024). Ripple rates, however, did not correlate significantly with reductions in hippocampal volume (r = − 0.23, p < 0.36).

Hippocampal 3D Reconstruction

Advanced MRI-based hippocampal surface mapping techniques were used to evaluate the distribution of local HA in patients with respect to the control subjects (Figure 2). The statistical probability or P maps shown in Figure 3 depict areas of significant atrophy in patients ipsilateral and contralateral to ictal onset. Probability values were color coded to distinguish hippocampal areas of significant atrophy (colored white and red) from areas that were not significantly different between patient and control groups (colored yellow, green, and blue). Hippocampal P maps ipsilateral to ictal onset revealed extensive atrophy distributed heterogeneously throughout the hippocampus on both superior and inferior surfaces (Figure 3, left column). Overall, these ipsilateral atrophic areas were significant after correcting for multiple comparisons (p<.01; see Methods). In contrast, contralateral P maps revealed a noticeably different distribution of atrophy, where HA was limited to a few circumscribed areas on the superior and inferior surfaces. Overall, the extent of atrophy depicted in contralateral P map was not significant (p=.07).

Figure 3
Probability or P maps depicting the distribution of statistically significant hippocampal atrophy in MTLE patients. Areas colored white and red indicate regions where patient hippocampi are significantly smaller than control hippocampi (ANOVA p < ...

In order to evaluate local HA in relation to microelectrode-recorded FR and Ripples, each microelectrode was registered to the respective ipsilateral or contralateral 3D hippocampus, and hippocampal areas within a 5 mm radius of each microelectrode tip were outlined. Figures 4A & 4B show the surface location of these local areas, and corresponding rates of FR and Ripple discharge that were recorded from each microelectrode in the depth. Rates of FR and Ripple discharge were color coded to distinguish microelectrodes with high (colored red) versus low (colored blue) rates of discharge, while areas colored gray reflect hippocampal regions greater than 5 mm from the tip of any microelectrode. Microelectrodes were more widely distributed in hippocampi ipsilateral (n=10 sites) to seizure onset than in contralateral hippocampi (n=8), which was a result of the placement of intracranial depth electrodes used in the clinical diagnostic evaluation of each patient’s seizure disorder. Ipsilateral FR maps revealed a predominance of higher FR rates, while lower FR rates were observed contralaterally (Figure 4A). In contrast, Ripple maps showed that higher Ripple rates were more common contralateral to seizure onset (Figure 4B).

Figure 4
HFO maps depicting rates of FR and Ripple occurrence (events per minute) in hippocampi of ten patients (18 microelectrode recording sites) with MTLE. Maps of sites ipsilateral ("Ipsi") to seizure onset reflect HFO data recorded from 10 microelectrodes, ...

Overall, significant atrophy (i.e. P map areas with P<0.05) accounted for 38% of total surface area within ipsilateral and contralateral hippocampal non-gray areas shown in Figure 4, while the remaining 62% was non-atrophic (i.e. P > 0.05). Rates of FR discharge were 94% higher in atrophic areas compared to non-atrophic areas (mean rate ± SE: 0.58 ± 0.008 per min vs. 0.30 ± 0.004 per min; p < 0.001). Furthermore, within these same atrophic regions higher rates of FR discharge correlated directly with greater atrophy (r = 0.65, p < 0.01). In contrast, rates of Ripple occurrence were not significantly different between atrophic and non-atrophic areas (0.069 ± 0.001 per min vs. 0.064 ± 0.001 per min; p < 0.60).

In a separate analysis that examined atrophy within a radius of 3 mm from each microelectrode, 43% of the total surface surrounding the microelectrodes was atrophic and 57% was non-atrophic. Rates of FR occurrence were 182% higher in atrophic areas than in non-atrophic areas (0.73 ± 0.02 per min vs. 0.26 ± 0.007 per min; p < 0.001), and correlated directly with magnitude of local atrophy (r = 0.72, p < 0.01). Analysis of Ripples did not reveal any difference in rates of occurrence between atrophic and non-atrophic areas (0.084 ± 0.002 per min vs. 0.095 ± 0.002 per min; p < 0.72).

Discussion

Data from the present study extend results from previous studies that examined anatomical disturbances associated with FR and Ripples, and also provide new information on the extent of HA in relation to FR and Ripples. Using microelectrode-recorded data in conjunction with hippocampal surface maps, we found that the occurrence of FR was significantly higher in regions of local atrophy than in non-atrophic areas. Furthermore, within atrophic regions, higher rates of FR discharge correlated with greater atrophy. In previous studies using kainic acid (KA) treated epileptic rats, FR were observed primarily within or adjacent to the KA-induced lesion,29, 52 but the extent of the lesion was not quantified, nor analyzed with respect to the rate of FR occurrence. Higher ratios of FR to Ripple discharge have been shown to correlate with greater CA3 neuron loss in pilocarpine-treated epileptic rats,22 and with smaller hippocampal volumes in patients with TLE,16 but these studies did not localize atrophy or cell loss specifically near the recording electrodes, and therefore do not indicate whether FR-generating sites coincide with local areas of HA. Unlike previously used techniques, 3D surface map-based methods retain information on the spatial distribution of HA, and, when combined with the registration of microelectrodes, allowed us to examine HA within several millimeters of the microelectrode. Results from our analysis suggest that local areas of atrophy are in close proximity to sites generating FR.

It is important to note that the maps shown in Figure 4 depict local hippocampal areas of FR and Ripple discharge adjacent to the microelectrodes, but likely do not represent the complete spatial distribution of FR and Ripple events. So although our data clearly show that FR occurrence is significantly higher within atrophic hippocampal areas, it is not clear how well these maps reflect the actual volume of tissue supporting the generation of FR. Voltage versus depth profile analysis using microelectrodes in KA-treated epileptic have estimated that 1 mm3 of tissue may be sufficient to support the generation of FR,21, 29 which may or may not be the actual volume of tissue generating FR. Recordings using macroelectrodes in patients with medically refractory epilepsy suggest FR-generating areas are larger,13, 20 but how much larger is not known. The present study found a stronger correlation between rates of FR occurrence and local HA when a more conservative 3 mm radius was used compared to 5 mm, suggesting that human hippocampal networks generating FR may occupy a smaller volume of tissue than those represented in Figure 4. Nevertheless, the strength of 3D surface map results, which were consistent with reductions in whole volume, suggest that the resolution of our technique was sufficient to detect local atrophy and characterize atrophy with respect to rates of FR discharge.

Analysis in the present study used group atrophy data and individual FR rates, except where microelectrode recording sites overlapped across patients, in which case averaged FR rates were used. It may be somewhat unexpected, therefore, to have found a significant correlation between local HA and FR rates given the potential variability in the distribution of atrophy between patients, and the low probability of capturing FR using fixed electrodes in any individual patient. However, a strength of the analysis is the detection of statistically significant atrophy that takes into consideration patient variability. Thus, the atrophy patterns we observed represent hippocampal areas consistently damaged across patients participating in this study. Furthermore, it is possible that in patients with MTLE who have had many years of uncontrolled seizures, there may be higher probability of capturing FR due to a larger area supporting FR or to a greater number of FR-generating sites. This explanation is consistent with studies using epileptic KA rats that showed that higher rates of seizure occurrence were associated with a greater the number of FR generating sites.52 Thus, our results may also indicate a greater density of FR-generating sites in atrophic epileptogenic areas, which typically are associated with substantial neuron loss.

Following hippocampal injury, neuron loss and synaptic reorganization are implicated as the morphological changes underlying the increased propensity for neuronal synchronization and the development of chronic spontaneous seizures in the KA rat.53 In KA-treated epileptic rats, greater neuron loss corresponds with greater axon sprouting, and likely greater synaptic organization and synchrony of discharges.54 It is thought that synaptic reorganization leads to the abnormal formation of small clusters of neurons that fire synchronous bursts of action potentials, and that FR represent the field potentials of these neuronal events.29 FR generating sites are widely and heterogeneously distributed, and when local inhibition decreases, evidence suggests that they can enlarge, coalesce, and synchronize.55 If a critical mass is reached, this highly synchronous bursting may result in seizure genesis. The mechanisms underlying seizure genesis proposed in these previous studies suggest interictal FR may be a surrogate marker of hippocampal regions capable of generating spontaneous seizures. Data in the present study, which derives from patients with hippocampal seizures, showed areas of significant HA were associated with higher rates of FR activity, and suggests local HA may be a marker of hippocampal epileptogenicity.

Some recent patient studies have found a strong association between Ripples and epileptogenic cortical and hippocampal areas,13, 14, 39 whereas other patient and animal studies have found reduced rates of Ripple discharge in epileptogenic hippocampus, which correlated with reduced Ammon’s horn neuron densities.16, 22 In the present study, we did not find a significant relationship between HA and rates of Ripple occurrence. It is possible that some of these differences may be due to the location of networks generating Ripples, e.g. hippocampus versus dentate gyrus or neocortex, or to the neuronal mechanisms underlying Ripple generation.56 It is likely, however, that not all Ripples arise from abnormal neuronal activity. A recent patient study on hippocampal and entorhinal cortex Ripples found firing patterns and phase relationships between Ripples and putative pyramidal cells and interneurons that were similar to neuronal discharge patterns that occur during Ripples generated in CA1 of naïve rats.17

We can conclude, however, there is a strong association between the distribution of local HA and FR discharges. FR may be a surrogate marker of areas associated with neuron loss and synaptic reorganization that contributes to hippocampal seizures in MTLE. And finally, 3D surface maps of local HA could, in turn, constitute a biomarker of FR generating sites, which could provide a non-invasive way to localize epileptogenic hippocampal regions.

Acknowledgement

This work was supported by the National Institutes of Health NS-02808 and NS-33310. The authors would like to thank Dr. Gary Mathern, Eric Behnke, Tom Karnesi, and Marina Barysheva for their assistance with this study.

Footnotes

Statement of Conflict of Interest: There are no conflicts of interest, real or apparent.

References

1. O'Keefe J, Nadel L. The hippocampus as a cognitive map. Oxford: Clarendon Press; 1978.
2. Suzuki SS, Smith GK. Spontaneous EEG spikes in the normal hippocampus. II. Relations to synchronous burst discharges. Electroencephalogr Clin Neurophysiol. 1988;69:532–540. [PubMed]
3. Buzsaki G, Horvath Z, Urioste R, et al. High-frequency network oscillation in the hippocampus. Science. 1992;256:1025–1027. [PubMed]
4. Bragin A, Engel J, Jr, Wilson CL, et al. High-frequency oscillations in human brain. Hippocampus. 1999;9:137–142. [PubMed]
5. Bragin A, Engel J, Jr, Wilson CL, et al. Electrophysiologic analysis of a chronic seizure model after unilateral hippocampal KA injection. Epilepsia. 1999;40:1210–1221. [PubMed]
6. Skaggs WE, McNaughton BL, Permenter M, et al. EEG sharp waves and sparse ensemble unit activity in the macaque hippocampus. J Neurophysiol. 2007;98:898–910. [PubMed]
7. Siapas AG, Wilson MA. Coordinated interactions between hippocampal ripples and cortical spindles during slow-wave sleep. Neuron. 1998;21:1123–1128. [PubMed]
8. Sirota A, Csicsvari J, Buhl D, Buzsaki G. Communication between neocortex and hippocampus during sleep in rodents. Proc Natl Acad Sci U S A. 2003;100:2065–2069. [PubMed]
9. Molle M, Yeshenko O, Marshall L, et al. Hippocampal sharp wave-ripples linked to slow oscillations in rat slow-wave sleep. J Neurophysiol. 2006;96:62–70. [PubMed]
10. Isomura Y, Sirota A, Ozen S, et al. Integration and segregation of activity in entorhinal-hippocampal subregions by neocortical slow oscillations. Neuron. 2006;52:871–882. [PubMed]
11. Axmacher N, Elger CE, Fell J. Ripples in the medial temporal lobe are relevant for human memory consolidation. Brain. 2008;131:1806–1817. [PubMed]
12. Jirsch JD, Urrestarazu E, LeVan P, et al. High-frequency oscillations during human focal seizures. Brain. 2006;129:1593–1608. [PubMed]
13. Urrestarazu E, Chander R, Dubeau F, Gotman J. Interictal high-frequency oscillations (100–500 Hz) in the intracerebral EEG of epileptic patients. Brain. 2007;130:2354–2366. [PubMed]
14. Worrell GA, Gardner AB, Stead SM, et al. High-frequency oscillations in human temporal lobe: simultaneous microwire and clinical macroelectrode recordings. Brain. 2008;131:928–937. [PMC free article] [PubMed]
15. Staba RJ, Wilson CL, Bragin A, et al. Quantitative analysis of high-frequency oscillations (80–500 Hz) recorded in human epileptic hippocampus and entorhinal cortex. J Neurophysiol. 2002;88:1743–1752. [PubMed]
16. Staba RJ, Frighetto L, Behnke EJ, et al. Increased Fast ripple to ripple Ratios Correlate with Reduced Hippocampal Volumes and Neuron Loss in Temporal Lobe Epilepsy Patients. Epilepsia. 2007;48:2130–2138. [PubMed]
17. Le Van Quyen M, Bragin A, Staba R, et al. Cell type-specific firing during ripple oscillations in the hippocampal formation of humans. J Neurosci. 2008;28:6104–6110. [PMC free article] [PubMed]
18. Jacobs J, LeVan P, Chander R, et al. Interictal high-frequency oscillations (80–500 Hz) are an indicator of seizure onset areas independent of spikes in the human epileptic brain. Epilepsia. 2008;49:1893–1907. [PMC free article] [PubMed]
19. Staba RJ, Wilson CL, Bragin A, et al. High-frequency oscillations recorded in human medial temporal lobe during sleep. Ann Neurol. 2004;56:108–115. [PubMed]
20. Khosravani H, Mehrotra N, Rigby M, et al. Spatial localization and time-dependant changes of electrographic high frequency oscillations in human temporal lobe epilepsy. Epilepsia. 2008 [Epub ahead of print] [PubMed]
21. Bragin A, Wilson CL, Engel J., Jr Chronic epileptogenesis requires development of a network of pathologically interconnected neuron clusters: a hypothesis. Epilepsia. 2000;41 Suppl 6:S144–S152. [PubMed]
22. Foffani G, Uzcategui YG, Gal B, Menendez de la Prida L. Reduced spike-timing reliability correlates with the emergence of fast ripples in the rat epileptic hippocampus. Neuron. 2007;55:930–941. [PubMed]
23. Bratz E. Ammonshornbefunde bei epileptikern. Arch Psychiatr Nervenkr. 1899;31:820–835.
24. Mathern GW, Pretorius JK, Babb TL. Quantified patterns of mossy fiber sprouting and neuron densities in hippocampal and lesional seizures. J Neurosurg. 1995;82:211–219. [PubMed]
25. Sommer W. Die Erkrankung des Ammonshorns als aetiologisches Moment der Epilepsie. Arch Psychiatr Nervenkr. 1880;10:631–675.
26. Babb TL, Brown WJ, Pretorius J, et al. Temporal lobe volumetric cell densities in temporal lobe epilepsy. Epilepsia. 1984;25:729–740. [PubMed]
27. Lin JJ, Salamon N, Dutton RA, et al. Three-dimensional preoperative maps of hippocampal atrophy predict surgical outcomes in temporal lobe epilepsy. Neurology. 2005;65:1094–1097. [PMC free article] [PubMed]
28. Ogren JA, Bragin A, Wilson CL, et al. Three-dimensional hippocampal atrophy maps distinguish two common temporal lobe seizure-onset patterns. Epilepsia. 2008 [Epub ahead of print] [PMC free article] [PubMed]
29. Bragin A, Mody I, Wilson CL, Engel J., Jr Local generation of fast ripples in epileptic brain. J Neurosci. 2002;22:2012–2021. [PubMed]
30. Bragin A, Wilson CL, Staba RJ, et al. Interictal high-frequency oscillations (80–500 Hz) in the human epileptic brain: entorhinal cortex. Ann Neurol. 2002;52:407–415. [PubMed]
31. Apostolova LG, Dinov ID, Dutton RA, et al. 3D comparison of hippocampal atrophy in amnestic mild cognitive impairment and Alzheimer's disease. Brain. 2006;129:2867–2873. [PubMed]
32. Bearden CE, Thompson PM, Dutton RA, et al. Three-Dimensional Mapping of Hippocampal Anatomy in Unmedicated and Lithium-Treated Patients with Bipolar Disorder. Neuropsychopharmacology. 2008;33:1229–1238. [PubMed]
33. Narr KL, Thompson PM, Szeszko P, et al. Regional specificity of hippocampal volume reductions in first-episode schizophrenia. Neuroimage. 2004;21:1563–1575. [PubMed]
34. Thompson PM, Schwartz C, Toga AW. High-resolution random mesh algorithms for creating a probabilistic 3D surface atlas of the human brain. Neuroimage. 1996;3:19–34. [PubMed]
35. Engel J., Jr . Fundamental mechanisms of human brain function: surgical treatment of epilepsy as an investigative resource. New York: Raven Press; 1987.
36. Engel J., Jr Surgery for seizures. N Engl J Med. 1996;334:647–652. [PubMed]
37. Ylinen A, Bragin A, Nadasdy Z, et al. Sharp wave-associated high-frequency oscillation (200 Hz) in the intact hippocampus: network and intracellular mechanisms. J Neurosci. 1995;15:30–46. [PubMed]
38. Chrobak JJ, Buzsaki G. High-frequency oscillations in the output networks of the hippocampal-entorhinal axis of the freely behaving rat. J Neurosci. 1996;16:3056–3066. [PubMed]
39. Bagshaw AP, Jacobs J, Levan P, et al. Effect of sleep stage on interictal high-frequency oscillations recorded from depth macroelectrodes in patients with focal epilepsy. Epilepsia. 2008 [Epub ahead of print] [PMC free article] [PubMed]
40. Collins DL, Neelin P, Peters TM, Evans AC. Automatic 3D intersubject registration of MR volumetric data in standardized Talairach space. J Comput Assist Tomogr. 1994;18:192–205. [PubMed]
41. Mai JK, Assheuer J, Paxinos G. Atlas of the human brain. San Diego: Academic Press; 1997.
42. Insausti R, Juottonen K, Soininen H, et al. MR volumetric analysis of the human entorhinal, perirhinal, and temporopolar cortices. AJNR Am J Neuroradiol. 1998;19:659–671. [PubMed]
43. Thompson PM, Apostolova LG. Computational anatomical methods as applied to ageing and dementia. Br J Radiol. 2007;80:S78–S91. Spec No 2. [PubMed]
44. Narr KL, Thompson PM, Sharma T, et al. Three-dimensional mapping of temporo-limbic regions and the lateral ventricles in schizophrenia: gender effects. Biol Psychiatry. 2001;50:84–97. [PubMed]
45. Thompson PM, Hayashi KM, De Zubicaray GI, et al. Mapping hippocampal and ventricular change in Alzheimer disease. Neuroimage. 2004;22:1754–1766. [PubMed]
46. Portney LG, Watkins MP. Foundations of clinical research: applications to practice. 3rd ed. Upper Saddle River, N.J.: Pearson/Prentice Hall; 2008.
47. Nichols TE, Holmes AP. Nonparametric permutation tests for functional neuroimaging: a primer with examples. Hum Brain Mapp. 2002;15:1–25. [PubMed]
48. Chiang MC, Reiss AL, Lee AD, et al. 3D pattern of brain abnormalities in Williams syndrome visualized using tensor-based morphometry. Neuroimage. 2007;36:1096–1109. [PMC free article] [PubMed]
49. Edgington ES. Randomization tests. 3rd ed. New York: M. Dekker; 1995.
50. Snijders TAB, Bosker RJ. Multilevel analysis : an introduction to basic and advanced multilevel modeling. Sage Publications: London ; Thousand Oaks, Calif.; 1999.
51. Zimmerman DW, Williams RH, Zurabo BD. Effect of nonindependence of sample observations on some parametric and nonparametric statistical tests. Communication in Statistics - Simulation and Computation. 1993;22:779–789.
52. Bragin A, Wilson CL, Engel J., Jr Spatial stability over time of brain areas generating fast ripples in the epileptic rat. Epilepsia. 2003;44:1233–1237. [PubMed]
53. Mathern GW, Cifuentes F, Leite JP, et al. Hippocampal EEG excitability and chronic spontaneous seizures are associated with aberrant synaptic reorganization in the rat intrahippocampal kainate model. Electroencephalogr Clin Neurophysiol. 1993;87:326–339. [PubMed]
54. Buckmaster PS, Dudek FE. Neuron loss, granule cell axon reorganization, and functional changes in the dentate gyrus of epileptic kainate-treated rats. J Comp Neurol. 1997;385:385–404. [PubMed]
55. Bragin A, Azizyan A, Almajano J, et al. Analysis of chronic seizure onsets after intrahippocampal kainic Acid injection in freely moving rats. Epilepsia. 2005;46:1592–1598. [PubMed]
56. Engel JJ, Bragin A, Staba R, Mody I. High-frequency oscillations: What is normal and what is not? Epilepsia. 2008 [Epub ahead of print] [PubMed]
57. Engel J., Jr . Surgical treatment of the epilepsies. 2nd ed. New York: Raven Press; 1993.
58. Mathern GW, Babb TL, Vickrey BG, et al. Traumatic compared to non-traumatic clinical-pathologic associations in temporal lobe epilepsy. Epilepsy Res. 1994;19:129–139. [PubMed]