A single case TMS study on an individual with Highly Superior Autobiographical Memory (HSAM)

Introduction We conducted the first TMS study investigating neural areas supporting a rare form of exceptional memory called HSAM (Mazzoni et al., 2019). Approximately 50 people (Patihis, 2015) in the world have this ability to remember nearly every day of their life, and studies show memory retrieval is extremely quick (Santangelo et al., 2022), detailed, and accurate (Parker et al., 2006). HSAM provides a unique angle to explore the neural basis of memory enhancement, and this knowledge could contribute to overcoming issues such as age-related memory decline or pathological neurodegeneration (Venneri et al., 2011). fMRI studies show HSAM memory access involves bilateral visual areas and the dorsolateral prefrontal cortex (DLPFC), but no study has explored the role of these brain areas with TMS. Methods DT is a healthy 32-year-old male who passed a rigorous exceptional memory screen. DT participated in 2 within-subject TMS experiments; he was shown blocks of 120 dates (e.g., 1st March 1999) and we assessed his ability to retrieve, for each, a verifiable event while online 10-Hz triple-pulse TMS was delivered (3 pulses with interstimulus interval of 100 ms). In Experiment 1, DT indicated when a memory was accessed, and classified retrieval type (direct or generative). TMS was initiated 500 ms after stimulus onset over left DLPFC, visual area (V1), or vertex. In Experiment 2, TMS was administered 100 ms after stimulus onset over V1 or vertex. After memory access and direct vs. generative choice, DT rated on a Likert scale memory point of view, ranging from 1 (1st person) to 7 (3rd person). Results In Experiment 1, for generative retrievals, during DLPFC stimulation, DT was slower accessing a memory for provided dates compared to vertex TMS, t(350) = 4.26, p < .001. Conversely, during V1 stimulation DT was faster accessing a memory for dates compared with vertex TMS, t(350) = −5.06, p < .001. No effects found for direct retrievals. In Experiment 2, no TMS site, or interaction effects were found on RTs. Post-hoc tests revealed, compared with vertex TMS, generative retrieval V1 stimulation caused memories to be rated as 3rd person to a higher degree, t(234) = 1.94, p = .05. In direct retrievals, V1 stimulation caused DT's memories to be rated more as 1st person, t(234) = −5.13, p < .001. Discussion We report the first single HSAM case TMS study. Our findings suggest V1 and DLPFC are key structures in HSAM access, both contributing in temporally and functionally different ways in direct vs. generative memory retrievals. DLPFC likely contributes to generative retrieval by assessing whether a memory meets search criteria, consistent with the reconstructive and hierarchic nature of such memory processes. Our findings suggest V1 modulates the content of direct and generative retrievals in DT, and links V1 to regulation of visual perspective during reliving. The performance enhancement observed in Experiment 1 during V1 stimulation could be an addition-by-subtraction mechanism, in which disruption of competing visual processes during generative retrievals leads to a more efficient search.