Decomposing Positive Memory Retrieval in Depression

Arkadiy Maksimovskiy, PhD

McLean Hospital
Decomposing Positive Memory Retrieval in Depression

Scientific Abstract


Background: Depression has been associated with poor memory for positive material. We attempted to replicate this result and examine whether it reflects slower evidence accumulation during retrieval by analyzing the data with the Hierarchical Drift Diffusion Model (HDDM).

Methods: Ninety-five adults filled out the Beck Depression Inventory II (BDI-II) and were divided into “High BDI” (HBDI; BDI-II>20, n = 49) and “Low BDI” (LBDI; BDI-II<6, n = 46) groups. They then completed an online memory study over two consecutive days. On the first day, they performed a visual oddball task in which neutral, negative, and positive pictures served as rare targets. On the second day, recognition memory for pictures was tested.

Results: A Group (HBDI, LBDI) x Stimulus (neutral, negative, positive) ANOVA on hit rates yielded a significant interaction. Relative to the LBDI group, the HBDI group showed lower hit rates for old positive pictures (p = 0.03), but not old neutral or negative pictures. There were no group differences in accuracy for new pictures (lures). The HDDM revealed slower evidence accumulation—lower drift rates—only during retrieval of old positive pictures in the HBDI group.

Conclusions: Consistent with prior findings, adults with elevated depressive symptoms showed poorer memory for positive stimuli. HDDM analyses revealed that this deficit reflects, at least in part, slow evidence accumulation at retrieval. These results thus identify a specific aspect of memory—the speed of evidence accumulation during retrieval—that might be targeted for intervention.

Live Zoom Session – April 21st

research Areas


Arkadiy Maksimovskiy, PhD, Christopher Okine, Daniel Dillon, PhD

Affiliated Website