The effect of Alzheimer’s disease and its progression on pyramidal cell gain and connectivity

Abstract

Alzheimer's disease affects our cognitive neurophysiology by loss of neurones, synapses and neurotransmitters. An improved mechanistic understanding of the human disease will facilitate new treatments. To this end, biophysically-informed dynamic causal models can support inferences around laminar and cell-specific disease effects from human non-invasive imaging. Based on pre-clinical models and effects of cholinesterase inhibitors, we hypothesised that Alzheimer's disease would affect the modulation of superficial pyramidal cell gain and extrinsic connectivity between pyramidal cells of different regions in hierarchical cognitive networks. Magnetoencephalography (MEG) was recorded during an auditory mismatch negativity task from healthy adults (n=14) and people with symptomatic Alzheimer's disease or mild cognitive impairment (n=45, all amyloid-biomarker positive) at baseline and after 16 months. Fourteen people from the symptomatic group had repeat magnetoencephalography at two weeks to assess test-retest reliability. Sensor-level data were analysed using t-tests of the mismatch negativity amplitude from 140ms to 160ms. The repetition effect was assessed with repeated-measures analysis of covariance, using the average evoked response in the mismatch negativity time window as the repeated measure. An absolute, intraclass correlation model of the test-retest data assessed mismatch negativity amplitude reliability. We then fitted dynamic causal models to the evoked responses over 500ms. Second-level parametric empirical Bayes across participants examined the effect of (1) group, patients vs controls, and (2) progression, baseline vs follow-up, on the model parameters reflecting pyramidal cell gain modulation and extrinsic connectivity. There was a significant effect of both disease and progression on the mismatch negativity amplitude (patients vs controls, T=-1.80, p=0.04; patient baseline vs follow-up, T=-2.72, p=.005), which had excellent reliability (ICC=0.95, p<.001). Parametric empirical Bayes revealed strong evidence (posterior probability>95%) that Alzheimer's disease reduced extrinsic connectivity and superficial pyramidal cell gain modulation, which was reduced further at follow up assessment. The mechanistic modelling confirmed the hypothesis that reduced superficial pyramidal cell gain modulation and extrinsic connectivity can explain the observed neurophysiological effect of Alzheimer's disease. This approach to non-invasive magnetoencephalography data may be used for experimental medicine studies of candidate treatments, and bridge clinical to preclinical models of drug efficacy.