Event‐related potential signatures of episodic memory decline predicting progression to Alzheimer’s disease in asymptomatic at‐risk subjects: a longitudinal study.

Filipa Raposo Pereira,Maximilien Chaumon,Bruno Dubois,Valentina La Corte,Nathalie George

DOI: https://doi.org/10.1002/alz.077073

2023-12-01

Abstract:Abstract Background The challenge in Alzheimer’s disease (AD) research ‐the world most prevalent dementia‐remains on finding preclinical markers that can predict progression to the disease. In this respect electrophysiological (EEG) measures seem promising, based on their sensitivity to the appearance of subtle episodic memory (EM) alterations, which become an AD risk‐factor when in the presence of B‐amyloid burden (AB+; AD‐biomarker) (Paitel et al., 2021; Olichney et al., 2011). We aimed to identify event‐related potential (ERP) signatures of AD‐linked EM alterations in elder cognitively normal subjects at‐risk for AD. Methods EEG data was collected longitudinally with an EGI 256‐channels cap, over 6 yearly sessions, in 15 controls (AB‐), 15 non‐converters (AB+), and 15 converters to prodromal‐AD (AB+; who converted during follow‐up but were only considered before diagnosis) from the INSIGHTpre‐AD cohort. We analysed the typical ERP response to a classic task of verbal memory recognition of old or new words (i.e. P3 and FN400 [0.25‐0.34 and 0.51‐0.57ms], P600 [0.62‐0.77ms], and monitoring [0.87‐1.04ms]). The new words were related to the old words (previously encoded) either in their number of repetitions (NR) or in their semantic content but presented once (NU). Source‐localization analysis of these data is currently underway. Results During follow‐up, the converters showed increased reaction time and decreased accuracy in word recognition when compared with the other groups. The converters significantly showed greater mean amplitude in the fronto‐parietal left and parieto‐occipital central regions typically linked to EM recollection, in the 0.25‐0.34 ms time‐window when compared with the controls, and in the 0.51‐0.57ms time‐window when compared with the non‐converters. Through our period of interest, old words activated mostly central‐posterior left regions (semantic networks/familiarity), NR words activated the frontal right region (i.e. novelty detection), and NU words frontal‐temporal right regions (semantic interference detection). Conclusion Our findings show amplitude differences associated with AD‐linked EM alterations years before individuals are diagnosed, representing potential ERP signatures of progression to AD in at‐risk individuals, while they remain cognitively normal. Anticipation of diagnosis contributes to the development of more effective tailor‐made treatments that potentially ease quality of life of patients and caregivers.

clinical neurology