Unique deficits in place coding across subfields of the hippocampus in a mouse model of temporal lobe epilepsy

Memory problems are comorbid with Temporal Lobe Epilepsy (TLE). Animal models of TLE reveal impairments in spatial firing fields of hippocampal place cells, providing a potential neural substrate for memory problems. Each subfield of the hippocampus carries out unique aspects of spatial memory, yet little is known about how individual subfields are perturbed. Here, we investigated the spatial coding properties of the three major subfields of the hippocampus.

Single unit recordings were made from CA1, CA3 and the dentate gyrus (DG) of mice (N = 10, 6M/4F) induced with epilepsy using the supra-hippocampal kainate model and in control mice injected with saline (N = 6, 3M/3F). Place cell activity was measured while mice foraged in highly familiar environments to assess basic place cell properties and in novel environments to assess remapping.

A lower percentage of cells were classified as place cells in CA1 of epileptic mice, whereas percentages were similar in CA3 and DG compared to control. Place fields of CA1 were less coherent, place fields of CA3 were less stable, and place fields in DG has smaller differences between in-field and out of field firing. All regions constructed new distinct maps within the first session of exposure to a novel environment, however new maps in CA3 trended toward instability. Significance These results point to specific deficits within subfields of the hippocampus, which may indicate that there are different cellular and network mechanisms at play. Such heterogeneity would be predicted to contribute differently to memory deficits. Key Points CA1 exhibits place map quality problems; CA1 has fewer place cells and those remaining have reduced spatial coherence CA3 exhibits place map stability problems; CA3 has lower spatial correlation within sessions and trends toward not forming stable new maps in a novel environment. DG exhibits reduced signal to noise; DG has smaller differences between in-field and out of field firing rates. Competing Interest Statement The authors have declared no competing interest. Footnotes Throughout, statistics have been replaced with generalized linear mixed models that account for animal identity when comparing between neurons recorded across groups. Given the shear number of comparisons (3 regions for every comparison), we have added a dedicated section in the supplement that presents every statistic presented in the paper. N, estimated means, and confidence intervals are all presented there. All figures show entire distributions (values for each neuron). There is a new table that makes it clear how many neurons were recorded in each animal and each region. The supplemental figures have been replaced with 5 brand new supplemental figures which present (1) criteria for classification of principal neurons vs interneurons, (2) an extended place field analysis, (2) oscillatory coherence between hippocampal regions, (4) the relationship between interictal spike rates and spatial coding properties, (5) and an extended histological analysis (including example histology for all KA mice included in the study (N=10). We have added an additional author, Dr. Meg Donahue, who performed additional analysis for the rebuttal.