Absolute number of three populations of interneurons and all GABAergic synapses in the human hippocampus.

The human hippocampus, essential for learning and memory, is implicated in numerous neurological and psychiatric disorders, each linked to specific neuronal subpopulations. Advancing our understanding of hippocampal function requires computational models grounded in precise quantitative neuronal data. While extensive data exist on the neuronal composition and synaptic architecture of the rodent hippocampus, analogous quantitative data for the human hippocampus remain very limited. Given the critical role of local GABAergic interneurons in modulating hippocampal functions, we employed unbiased stereological techniques to estimate the density and total number of three major GABAergic cell types in the male and female human hippocampus: parvalbumin (PV)-expressing, somatostatin (SOM)-positive, and calretinin (CR)-positive interneurons. Our findings reveal an estimated 49,400 PV-positive, 141,500 SOM-positive, and 250,600 CR-positive interneurons per hippocampal hemisphere. Notably, CR-positive interneurons, which are primarily interneuron-selective in rodents, were present in humans at a higher proportion. Additionally, using 3-dimensional electron microscopy, we estimated approximately 25 billion GABAergic synapses per hippocampal hemisphere, with PV-positive boutons comprising around 3.5 billion synapses, or 14% of the total GABAergic synapses. These findings contribute crucial quantitative insights for modeling human hippocampal circuits and understanding its complex regulatory dynamics. Understanding the operating principles of cortical networks is a major goal in neuroscience, requiring a detailed description of cortical neurons, including their proportions, densities, and, most critically, their absolute numbers. Such precise quantification is essential for constructing predictive computational models of neuronal circuits. In this study, we present the density and total number of three major, non-overlapping GABAergic cell groups across each layer of the human hippocampus. Additionally, we quantify the total number of all GABAergic boutons and synapses, as well as separately quantifying those originating from parvalbumin (PV)-positive cells within each of the 15 hippocampal layers. This dataset provides a foundational framework for accurately modeling hippocampal inhibition at the cellular level.