Phase-locking characteristics of limbic P3 responses in hippocampal sclerosis

Abstract

Amplitudes of the P3 recorded invasively from the medial temporal lobe (MTL-P3) have been reported to be reduced on the side of a mediotemporal epileptogenic focus. This reduction has been attributed to the massive cell loss within the hippocampus associated with hippocampal sclerosis. It has remained unclear how functional connectivity between the hippocampus and rhinal cortex, as well as within the hippocampus, is altered in hippocampal sclerosis. To investigate this issue, we analyzed to what extent stimulus-related phase-locking and power changes within the low-frequency range (2–30 Hz) and within the gamma band (32–48 Hz), as well as rhinal–hippocampal phase synchronization contribute to the averaged MTL-P3 potentials. Event-related responses were recorded via bilateral depth electrodes in epilepsy patients with unilateral hippocampal sclerosis, who performed a visual oddball experiment. On the contralateral (nonsclerotic) side, successful target detection was associated with an increase of power and phase locking of hippocampal activity in both the low-frequency range and in the gamma range. Besides, there were rhinal–hippocampal synchronization enhancements in the theta and gamma range. On the ipsilateral (sclerotic) side, the event-related power increase in the low-frequency range had almost disappeared, a finding likely to be explained by the loss of principle neurons. However, low-frequency phase-locking, rhinal–hippocampal synchronization, as well as event-related power changes in the gamma range persisted ipsilaterally, although there were differences in temporal and spectral characteristics. These findings support the hypothesis that functional connectivity between hippocampus and rhinal cortex, as well as intrahippocampal connectivity, are partially preserved in hippocampal sclerosis.

Introduction

The so-called limbic or medial temporal lobe P3 (MTL-P3), similar to the P3 recorded from the scalp, is the most prominent component of the event-related response elicited by rare target stimuli in auditory or visual ‘oddball’ paradigms. In contrast to the multiple sources underlying the widespread surface P3, the MTL-P3 (e.g., Halgren et al., 1998), which actually represents a negative deflection, appears to be generated locally within the hippocampus (e.g., Grunwald et al., 1999a, Halgren et al., 1980). Although its functional significance is not yet clear, the MTL-P3 may be associated with the hippocampal contribution to tasks of updating or closure of a context within working memory, the assumed functional correlate of the P3 recorded from the scalp (Donchin and Coles, 1988, Verleger, 1998). It has been reported that MTL-P3 amplitudes are reduced unilaterally on the side of the primary epileptogenic focus in temporal lobe epilepsy (e.g., Grunwald et al., 1995, Meador et al., 1987). This reduction of MTL-P3 amplitudes has been attributed to the most common morphological alteration of the hippocampus in temporal lobe epilepsy called hippocampal sclerosis (Puce et al., 1989), which is defined by a massive cell loss in the areas CA1, CA3 and the hilus of the hippocampus (Wyler et al., 1992). This damage results mainly in the loss of principal rather than inhibitory neurons (Babb and Brown, 1986, Blümcke et al., 1996, Wittner et al., 2001, Wittner et al., 2002). Consequently, in hippocampal sclerosis, MTL-P3 amplitudes on the sclerotic (ipsilateral) side have been found to be reduced to less than 50% of the amplitudes on the unaffected (contralateral) side (Grunwald et al., 1999a).

Based on the previous finding of reduced MTL-P3 amplitudes in hippocampal sclerosis, the aim of the present study was to determine the alterations of functional connectivity between the hippocampus and the parahippocampal region including the rhinal cortex, as well as changes of intrahippocampal connectivity occurring in hippocampal sclerosis. Event-related responses were recorded via depth electrodes in epilepsy patients with hippocampal sclerosis performing a visual oddball experiment. Apart from the conventional evaluation of event-related components, the contribution of phase-locking and power changes to the P3 responses within the low-frequency range (2–30 Hz) and within the gamma band (32–48 Hz), as well as phase synchronization between rhinal cortex and hippocampus were analyzed.

Stimulus-related power increases are thought to correspond to the event-related activation of neural assemblies distinct from those involved in ongoing background dynamics (e.g., Lopes da Silva, 1993, Mangun, 1992). A stimulus-related power enhancement, which is accompanied by an increase of inter-trial phase locking in the same time and frequency range, is probably related to a stimulus-locked evoked response (e.g., Klimesch et al., 2004, Shah et al., 2004). On the other hand, a phase-locking increase, which is dissociated from power enhancement in the time and frequency domain, is thought to characterize phase reset of ongoing oscillatory activity (e.g., Jansen et al., 2003, Makeig et al., 2002).

Since loss of principle neurons should result in decreased activation of neural assemblies, we hypothesized that event-related power increases in the low-frequency range should be reduced in hippocampal sclerosis. Hippocampal event-related gamma activity, on the other hand, is thought to depend on the integrity of intrahippocampal networks of interneurons, such that gamma oscillations arise either by glutamatergic/GABAergic synaptic interplay of interneurons and principal neurons (Csicsvari et al., 2003, Traub et al., 1996, Traub et al., 1999, Whittington et al., 1995) or by interneuronal gap junctional coupling (Buhl et al., 2003). Histological studies suggest, that, in spite of significant loss of principal neurons, intrahippocampal connectivity may be partly preserved in hippocampal sclerosis (Babb and Brown, 1986, Blümcke et al., 1996, Wittner et al., 2001, Wittner et al., 2002). Thus, we further hypothesized that power changes in the gamma range observed contralaterally would also be found on the ipsilateral side.

It has been speculated that hippocampal phase locking depends on external input from the parahippocampal region, in particular the entorhinal cortex (Borisyuk and Hoppensteadt, 1999, Buzsáki, 2002, Soltesz et al., 1993, Yamaguchi, 2003). Because of the evidence for a partially sustained rhinal–hippocampal connectivity in hippocampal sclerosis (e.g., Spencer and Spencer, 1994), we finally hypothesized that phase-locking changes in the low-frequency and in the gamma range, as well as rhinal–hippocampal phase synchronization detected on the contralateral side should persist ipsilaterally.