Memory-enhancing amygdala stimulation elicits gamma synchrony in the hippocampus

Abstract

Activation of the amygdala either during emotional arousal or by direct stimulation is thought to enhance memory in part by modulating plasticity in the hippocampus. However, precisely how the amygdala influences hippocampal activity to improve memory remains unclear. In the present study, brief electrical stimulation delivered to the basolateral complex of the amygdala (BLA) following encounters with some novel objects led to better memory for those objects 1 day later. Stimulation also elicited field-field and spike-field CA3-CA1 synchrony in the hippocampus in the low gamma frequency range (30-55 Hz), a range previously associated with spike timing and good memory. In addition, the hippocampal spiking patterns observed during BLA stimulation reflected recent patterns of activity in the hippocampus. Thus, the results indicate that amygdala activation can prioritize memory consolidation of specific object encounters by coordinating the precise timing of CA1 membrane depolarization with incoming CA3 spikes to initiate long-lasting spike-timing dependent plasticity at putative synapses between recently active neurons.

(c) 2015 APA, all rights reserved).

Figures

Figure 1

Overview of procedure: memory task, example histology, and memory performance. A, The illustration depicts a rat receiving brief unilateral electrical stimulation of the BLA immediately after exploring a novel object during the Study Phase as well as the rat encountering a duplicate of the same object again during the object recognition memory Test Phase. B, The schematic shows a full experimental procedure in which a rat encountered one object from each of three conditions on each trial while completing clockwise laps on a circular track. Stimulation objects (red, “S”) were followed by brief electrical stimulation to the BLA (denoted by a red star) immediately following the offset of exploration during the Study. No Stimulation objects (blue, “O”) were not followed by stimulation. New objects (green, “N”) were replaced by novel objects on the test. Object recognition memory was assessed on either the Immediate Test or the 1-Day Test. C, The left panel shows the postmortem marking lesion of a stimulating electrode localized to the BLA (dashed line) in a section stained for acetylcholinesterase (lateral nucleus, L; basal nucleus, BN; accessory basal nucleus, AB). Electrode tips were localized to the left and right BLA in all rats included for analysis. On the right, sections were stained with cresyl violet to facilitate localization of recording tetrodes in the pyramidal layer of CA1 (upper) and CA3 (lower) in the hippocampus. D, Rats remembered Stimulation objects better than No Stimulation objects on the 1-Day Test (p < 0.05). There was no difference in memory for objects on the Immediate Test. The dashed line indicates chance performance. Error bars show the SEM (n = 7).

Figure 2

Gamma synchrony between CA3 and CA1 increased during novel object exploration. A, The top panel shows a moving time window coherogram and indicates that local field potentials in CA3 and CA1 of the hippocampus synchronized in the low gamma range following the onset of novel object exploration (at 0 s). The panel on the bottom left shows the difference in coherence between the first 2 s of object exploration (Exploration) and the 2 s prior to exploration (Baseline), and the panel on the bottom right shows only pixels that exceeded statistical significance (p < 0.01) when evaluated using a cluster-based randomization procedure (see Method). B, The top panels show mean coherence (left) and phase coherence (right) during baseline (blue) and exploration (red), and the bottom panels show the difference between exploration and baseline. Frequency bands that exceed statistical significance when evaluated using a cluster-based randomization procedure are highlighted in purple (p < 0.01). Shaded bands above and below the mean indicate SEM (n = 5).

Figure 3

Evoked field potentials in the hippocampus from first, second, and third pulse trains of BLA stimulation. Traces of local field potentials recorded from CA3 and CA1 during the first (top panel), second (middle panel), and third (bottom panel) 4-pulse train of ipsilateral (left column in each panel) or contralateral (middle column in each panel) 50 Hz stimulation were averaged within rats and then averaged across rats. Each pulse of ipsilateral but not contralateral BLA stimulation evoked, about 24 ms later, prominent negative potentials in CA3 (bottom row in each panel) and smaller but reliable positive potentials in CA1 (top row in each panel). For each 4-pulse train, the stimulation artifact visible for both ipsilateral and contralateral stimulation was largely eliminated when the mean ipsilateral-contralateral difference was calculated for each rat and then averaged (right column). The magnitude of the evoked response in both CA3 and CA1 to the first ipsilateral pulse from each train increased from the first to second 4-pulse train but did not appreciably increase further from the second to third 4-pulse train. Inverted triangles indicate each pulse. Shaded bands above and below the mean indicate SEM in each panel (n = 5).

Figure 4

Gamma synchrony between CA3 and CA1 increased during stimulation of the BLA. A, A prominent low gamma frequency oscillation is observed in the cross-correlation between CA3 and CA1 during ipsilateral but not contralateral BLA stimulation. Subtracting the contralateral results from the ipsilateral results removed the stimulation artifact but left the low gamma oscillation largely unchanged. Shaded bands above and below the mean indicate SEM (n = 5). B, A moving time window coherogram shows that local field potentials in CA3 and CA1 of the hippocampus synchronized in the low gamma range to a much greater extent during ipsilateral stimulation as compared to contralateral stimulation. The bottom left panel shows the ipsilateral-contralateral difference in coherence, and the bottom right panel shows only pixels from the same contrast that were statistically significant (p < 0.01) when evaluated using a cluster-based randomization procedure. C, The top panels show mean coherence (left) and phase coherence (right) during ipsilateral stimulation (red) and contralateral stimulation (blue). The bottom panels show the difference between ipsilateral and contralateral stimulation, and frequency bands that were statistically significant (p < 0.01) when evaluated using a cluster-based randomization procedure are highlighted in purple. Note: the response visible in the coherence plots at around 90-100 Hz reflects a harmonic introduced by the Fourier transform of the response in the low gamma range. Shaded bands above and below the mean indicate SEM (n = 5).

Figure 5

Gamma synchrony between CA3 and CA1 increased during stimulation of the BLA at 80 Hz. A moving time window coherogram shows that local field potentials in CA3 and CA1 of the hippocampus synchronized in the low gamma range as well as in the 80 Hz range to a much greater extent during ipsilateral stimulation as compared to contralateral stimulation. The bottom left panel shows the ipsilateral-contralateral difference in coherence, and the bottom right panel shows only pixels from the same contrast that were statistically significant (p

Figure 6

Gamma synchrony between CA3 pyramidal neuron spikes and CA1 local field potentials increased during stimulation of the BLA. A, Histograms in both standard (top row) and circular (bottom row) format show distributions of CA3 spike phases of low gamma (30-55 Hz) oscillations in the downstream CA1 field potential during ipsilateral and contralateral BLA stimulation. The white arrows on the circular histograms show the mean resultant vector, and the mean resultant length (MRL) of that vector indicates the strength of phase preference. B, To determine whether the MRL was significantly greater for ipsilateral as compared to contralateral BLA stimulation, the ipsilateral-contralateral difference was obtained from 2000 random shuffles of the data and a distribution of these random MRL difference was plotted for spike-field phases for CA3-CA1 (left), CA3-CA3 (middle), and CA1-CA1 (right). The red asterisk on each plot indicates the actual difference, and vertical dashed lines indicate the threshold for statistical significance (see Method). The ipsilateral-contralateral low gamma MRL difference was significantly greater than chance for CA3-CA1 and CA3-CA3 spike-field phases.

Figure 7

Hippocampal pyramidal neuron firing rates did not differ between ipsilateral and contralateral BLA stimulation but ipsilateral BLA stimulation was more associated with recent patterns of activity of hippocampal neurons. A, Perievent rasters show example spiking activity from active CA1 and CA3 pyramidal neurons before and during one ipsilateral and one contralateral BLA stimulation. The numbers above each raster indicate the multi-dimensional distance between spike patterns in each epoch, a measurement of how similar (lower = more similar) the pattern of activity was between the stimulation period (0 to +1 s) and the 1 s baseline (−1 to 0 s) preceding stimulation. B, The mean firing rate of each CA1 and CA3 pyramidal cell was averaged within condition and then across rats (n=7) during stimulation. Control stimulation was defined as the 1 s period after exploration of New objects when the rat would have received stimulation on that trial had the object been a Stimulation object. The results show that stimulation did not impact the firing rate of pyramidal neurons in the hippocampus (see Results; all ps > 0.1). Units that were not active during object exploration, stimulation, or the sham stimulation were not included in analyses. Error bars show the SEM across rats (n = 7). C, The mean baseline-stimulation ensemble distance for combined CA1 (n=189) and CA3 (n=163) pyramidal neurons is plotted for ipsilateral and contralateral stimulation for 5 rats each with > 20 pyramidal neurons (ns = 39, 42, 72, 87, and 112) and indicates greater similarity between stimulation and baseline periods for the ipsilateral BLA stimulation (p < 0.01).