Deep Brain Stimulation (DBS) is an FDA-approved treatment for Parkinson's Disease (PD), for which the medical workup includes routine pre- and post- operative neuropsychological assessment to determine potential surgical cognitive risk. Existing research suggests that cognitively normal individuals experience good cognitive outcome, whereas those with pre-existing cognitive deficits are prone to accelerated cognitive decline post-DBS. The goal of this study is to identify characteristics that determine which individuals with PD are at risk for accelerated post-DBS cognitive loss, and to characterize the nature of the decline in this population.

We conducted a retrospective chart review of PD- DBS patients who completed their DBS workup and surgery at Mount Sinai Hospital NYC between 2015 and 2022. Non-English speakers were excluded from this study due to small sample size and use of a neurocognitive battery different from that of English speakers. Using repeated measures t-tests, chi square, and regression analyses, we explored variables related to disease (e.g., duration, L-Dopa burden, DBS target), socio-demographic background (e.g., age onset, current age, education), assessment modality (telehealth vs in-office), neurocognitive performances (e.g., WMS-IV Logical Memory (LM), HVLT-R, WASI-II Matrix & Similarities, WAIS-IV Digit Span), and cognitive diagnosis (amnestic vs non-amnestic MCI) for all individuals in the sample. At the individual level, we utilized Reliable Change Indices (RCI) to identify clinically significant cognitive differences from pre- to post-DBS exam. We considered LM- Delayed Recall (LMDR) as a proxy for memory loss, as this cognitive function is expected to remain generally unchanged post PD-DBS. Therefore, decline on this measure in the first year after DBS could indicate a change in global memory function and possible evidence of accelerated postoperative decline.

Of 65 charts reviewed, 44 patients were native English-speaking and included in our analyses. At the group level, there were no significant differences in disease characteristics, socio-demographic variables, or cognitive classification between those who declined versus those who did not decline on LMDR. Regression statistics for predictors of cognitive decline also were non-significant. Of the eight individuals who declined on LMDR, one patient declined on a total of one neuropsychological measure, four declined on a total of two measures, two declined on a total of three measures, and one declined on a total of four measures. Two of these eight individuals had a diagnosis that changed to amnestic MCI based on concomitant interval history of ADL compromise. Of these two individuals, one declined in two tests and the other declined in four tests. Six of the eight individuals who declined also showed abnormalities in their imaging with either edema or hemorrhage.

Our analysis is unique in that we explored cognitive decline at both the group and individual levels. Despite this, we did not find predictors of post-DBS cognitive decline. Further detailed analysis of additional post-operative factors that might play a greater role in our understanding of this phenomenon is warranted. This said, our data do support that the majority of individuals with non-amnestic MCI did not decline cognitively.