Model-driven tACS in Alzheimer’s disease with personalized hybrid brain models

Abstract

Abstract Animal models suggest that 40 Hz induced gamma oscillations reduce Aβ and p-Tau deposition, which is considered central to Alzheimer’s disease (AD) pathophysiology. Transcranial alternating current stimulation (tACS) has been proposed as a feasible option to induce gamma in AD patients and thus ease the cognitive burden of the disease. Here we investigate the potential of personalized whole-brain models to drive the stimulation in AD patients. We build a hybrid brain model that combines the (i) anatomical information, i.e., a parcellated brain, with subcortical parcels and their connections inferred from dMRI, (ii) functional data, i.e. fMRI-based functional connectivity and (iii) physiology based on Laminar Neural Mass Models (NMM). The NMM, personalized further with TMS-EEG data, can represent fast (gamma, 40-80 Hz) and slow (alpha/beta, 4-22 Hz) oscillations. Pathological disruptions have been incorporated in the most affected parcels based on FDG PET SUVR, which result in a decrease of gamma power in said parcels, consistent with findings in EEG and TMS-EEG. We have explored stimulation strategies by targeting small networks of parcels and specific regions, such as the prefrontal cortex (PFC). We show that the interplay of resonance phenomena and network effects can facilitate gamma entrainment and recover part of the gamma-power loss. This study is the first step into building model-driven optimization strategies for non-invasive stimulation in AD. Research Category and Technology and Methods Translational Research: 8. Transcranial Alternating Current Stimulation (tACS) Keywords: tACS, alzheimer's disease, computational neuroscience, resonance phenomena