Alzheimer’s disease (AD) is a progressive neurodegenerative disease that affects millions of people. Repetitive transcranial magnetic stimulation (rTMS) is a noninvasive stimulation typically used in psychiatric conditions such as depression and anxiety. rTMS works by using an electric current to generate a transient magnetic field, depolarizing neurons in a target region and creating lasting changes in brain connectivity via synaptic plasticity. Patients with AD show disruptions in the Default Mode Network (DMN), a network of brain regions typically active during rest and crucial for memory consolidation. We hypothesize that strengthening the DMN through rTMS targeted at the left Brodmann 8AV region, selected for being an easily accessible node of the DMN, will improve memory in AD patients. To test this hypothesis, we are conducting a single-blind, single-arm, randomized cross-over trial of rTMS on early-stage AD patients over a 12 week period with week 1 where we scan for the 8AV region via MRI, during week 3 and 8 being the placebo or treatment week. We measure our primary outcome of the participants’ speed of forgetting —a novel index of memory function—through an individualized, adaptive memory test. To eliminate potential confounding variables, we also measure depression and anxiety symptoms during the 1st, 8th and 12th week of the study. Additionally, functional MRI scans will be analyzed for potential structural or functional differences caused by treatment. Preliminary results from our initial participants have shown promising improvements, and we are hopeful that similar outcomes will be observed in the remaining participants. Successful results would provide a novel target for AD treatment using rTMS, and support further investigation of rTMS as a viable treatment option.

Overgeneral Autobiographical Memory (OGM) is a common symptom of depression and Post-Traumatic Stress Disorder (PTSD). Instead of remembering specific details, individuals with OGM can only describe past events from their lives in general terms. The "trauma hypothesis" suggests that OGM emerges because individuals suffering from PTSD tend to reduce the number of details they retrieve about their memories to avoid remembering their trauma. However, this hypothesis does not fully explain how this avoidance is learned, or why avoidance spreads from traumatic memories to all autobiographical memories. To this end, we propose a computational model of OGM that integrates theories of memory retrieval and trauma with reinforcement learning. In this model, multiple episodic memories are nodes in an interconnected network, and memories are retrieved when visiting that node in the network. The more nodes that are visited, the more detailed that autobiographical recall will be. On the other hand, visiting more nodes comes with an increased risk of encountering a traumatic memory, which comes with negative emotional valence. The decisions about whether to visit another node or terminate the retrieval process are made using reinforcement learning, which takes actions based on predicted outcomes. By obtaining a greater understanding of OGM through this model, we hope to improve treatments for PTSD that specifically targets its effects on memory.