Schizophrenia genome-wide association studied with computer simulation: gamma oscillations and information flow in CA3

TitleSchizophrenia genome-wide association studied with computer simulation: gamma oscillations and information flow in CA3
Publication TypeConference Paper
Year of Publication2016
AuthorsSherif, M. A., Neymotin S. A., & Lytton WW.
Conference NameSociety for Neuroscience 2016 (SFN '16)
KeywordsSFN, Society for Neuroscience
Abstract

Brain oscillations, which play roles in memory and attention, are abnormal in schizophrenia (SCZ). Genome wide association studies (GWAS) in SCZ have identified 108 loci of possible mutations. We looked at two putative mutations, using multiscale modeling of hippocampal area CA3, to explore their effects on oscillations and information flow. Mutations at these 2 loci will alter network activity through changes in cell conductances: 1. HCN1 – coding the channel mediating the h-current (Ih); 2. GRIN2A – coding subunit 2A of the NMDA-type receptor (NMDAR). Our network model consisted of pyramidal neurons (PYR), basket cells (BAS) and Oriens Lacunosum Moleculare (OLM) interneurons. BAS and OLM also received inhibitory input from medial septum (MS). We varied the magnitude of Ih and NMDAR conductances (gh and gNMDAR respectively) at sites on each cell type (Table 1) and in combinations. We evaluated the simulations for increase in gamma oscillation, a putative signature for SCZ. We also looked at decreased theta power as a possible secondary marker. Using multiple simulation wirings and drive, we found consistent gamma increase with increased Ih conductance at BAS, or with decreased gNMDAR conductance at OLM. These oscillatory changes were also associated with reduction of information transfer from CA3 input (mossy fibers of the trisynaptic pathway and perforant pathway) to output (Schaffer collaterals) as measured by normalized transfer entropy (nTE). We predict that the increase in gamma power seen in SCZ would be a consequence of HCN1 mutations that increase gh primarily on BAS, and/or GRIN2A mutations that decrease gNMDAR primarily on OLM cells. Our multiscale modeling of CA3 showed how 2 GWAS loci can produce oscillatory signatures of SCZ that would relate to alterations in information flow with implications for cognitive dysfunction. We propose that both HCN1 and GRIN2A are part of the same ``clinical pathway'' involved in generating oscillations (a potential biomarker), and producing cognitive impairment.