Thapaliya, B. | bioRxiv Science

bioRxiv2026

Film Recall Reveals Intact Event Memory but Impaired Sequence Memory in Temporal Lobe Epilepsy Patients

BACKGROUND AND OBJECTIVESPatients with epilepsy (PWE), especially temporal lobe epilepsy (TLE), experience impaired memory for personally experienced events. However, current assessments of episodic memory are limited in their ecological validity with a potential to miss detection of subtle cognitive decline. We conducted an exploratory study to determine whether a naturalistic film-viewing task with open-ended spoken recall could detect memory differences between TLE patients and healthy controls (HCs). METHODSTLE patients (ages 18-60, fluent in English, not legally blind) were recruited from a Level 4 Epilepsy Center (2018-2024). TLE diagnosis was based on seizure semiology, MRI Brain, and EEG. TLE patients scored [≥]22/30 on the Montreal Cognitive Assessment (MOCA); HCs scored [≥]26/30. Subjects watched 6 short films and then freely recalled film details. Spoken responses were recorded, transcribed, segmented, and scored for film- and event-level recall. Recall order was assessed using the Damerau-Levenshtein distance. Semantic and causal centrality were quantified using sentence embeddings and rater-identified causal links, respectively. Beta regression with cluster-robust standard errors assessed group and centrality effects on recall probability. Beta regression evaluated the influence of age, MOCA, and testing platform on sequence recall error. RESULTSWe recruited 51 subjects (27 TLEs; 24 HCs, 70.1% F, mean 29.9 {+/-}8.3 years). TLE patients and HCs showed similar recall of films (HC 89% {+/-}11% vs TLE 88% {+/-}18%, p = 0.54), coarse events (HC 50% {+/-}16% vs TLE 44% {+/-}18%, p = 0.19) and fine events (HC 25%{+/-}10% vs. TLE 22%{+/-}12%, p=0.17). Both groups recalled high causal centrality events better. For coarse event sequence recall, TLE patients showed a numerical trend toward greater sequence errors compared to HCs (HC 10.8% {+/-} 10.5% vs. TLE 19.5% {+/-} 18%, p = 0.06), although this difference did not reach statistical significance. However, TLE patients showed significantly greater fine event sequence errors at recall than HCs (HC 15% {+/-}13% vs 23% {+/-}18%, p = 0.02, Hedges g = 0.85, Cliffs {delta} = 0.51), with RTLE demonstrating more sequence errors than HCs (15%{+/-}13 vs. 29%{+/-}21% p = 0.021) Age, education, MOCA, and performance on standard verbal and visual memory tasks were unrelated to film, event, and sequence recall performance. DISCUSSIONWe demonstrate that a short film task with spontaneous spoken recall can identify sequence memory impairment in TLE patients despite intact film- and event-level recall. Sequence memory may represent a subtle manifestation of memory impairment that is not detected by standard cognitive testing. Key PointsO_LIWe asked whether a naturalistic film recall task could detect episodic memory impairment in a temporal lobe epilepsy cohort. C_LIO_LIPatients with temporal lobe epilepsy showed comparable film and event recall compared to healthy controls but were found to have impaired sequence memory. C_LIO_LISequential memory for temporal order is an overlooked aspect of episodic memory that may detect subtle memory decline. C_LI

neuroscience↗

bioRxiv2024

Efficient Federated Learning for distributed NeuroImaging Data

Recent advancements in neuroimaging have led to greater data sharing among the scientific community. However, institutions frequently maintain control over their data, citing concerns related to research culture, privacy, and accountability. This creates a demand for innovative tools capable of analyzing amalgamated datasets without the need to transfer actual data between entities. To address this challenge, we propose a decentralized sparse federated learning (FL) strategy. This approach emphasizes local training of sparse models to facilitate efficient communication within such frameworks. By capitalizing on model sparsity and selectively sharing parameters between client sites during the training phase, our method significantly lowers communication overheads. This advantage becomes increasingly pronounced when dealing with larger models and accommodating the diverse resource capabilities of various sites. We demonstrate the effectiveness of our approach through the application to the Adolescent Brain Cognitive Development (ABCD) dataset.

neuroscience↗

bioRxiv2023

Replication and Refinement of Brain Age Model for adolescent development.

The discrepancy between chronological age and estimated brain age, known as the brain age gap, may serve as a biomarker to reveal brain development and neuropsychiatric problems. This has motivated many studies focusing on the accurate estimation of brain age using different features and models, of which the generalizability is yet to be tested. Our recent study has demonstrated that conventional machine learning models can achieve high accuracy on brain age prediction during development using only a small set of selected features from multimodal brain imaging data. In the current study, we tested the replicability of various brain age models on the Adolescent Brain Cognitive Development (ABCD) cohort. We proposed a new refined model to improve the robustness of brain age prediction. The direct replication test for existing brain age models derived from the age range of 8-22 years onto the ABCD participants at baseline (9 to 10 years old) and year-two follow-up (11 to 12 years old) indicate that pre-trained models could capture the overall mean age failed precisely estimating brain age variation within a narrow range. The refined model, which combined broad prediction of the pre-trained model and granular information with the narrow age range, achieved the best performance with a mean absolute error of 0.49 and 0.48 years on the baseline and year-two data, respectively. The brain age gap yielded by the refined model showed significant associations with the participants information processing speed and verbal comprehension ability on baseline data.

neuroscience↗