Keynotes | K&Ts | GACs | Talks | Posters | Search

Poster E in Poster Session E: Thursday, August 6, 10:30 am – 12:15 pm, Kimmel Center, Shorin & Rosenthal Rooms

Normative modelling characterises individual brain heterogeneity and improves cognitive prediction in chronic moderate-severe traumatic brain injury

Gavin Gan¹, Jake Mitchell¹, Amber Ayton¹, Alexia Samiotis¹, Amelia Hicks¹, Meng Law¹, Sandy Shultz^1,2,3, Stuart McDonald^1,3, Rachael Knott¹, Jennie Ponsford¹, Gershon Spitz¹; ¹Monash University, ²The Alfred Hospital, ³Vancouver Island University

Presenter: Gavin Gan

Cognitive outcomes after traumatic brain injury (TBI) vary widely between individuals, yet most research relies on group averages that obscure this heterogeneity. This study examined whether normative modelling, a person-centred approach that quantifies how individual brain measures deviate from population expectations, can better characterise brain differences and improve prediction of cognition in chronic moderate-to-severe TBI. Individual brain deviation scores were computed using normative models trained on 58,046 brain scans, adjusting for age, sex, and site effects across cortical and subcortical regions and compared with raw brain measures. Group-level analyses suggested widespread brain differences, but normative modelling showed that most individuals fell within typical ranges, highlighting marked person-to-person variability. Models using normative brain measures predicted cognitive performance more accurately than models using raw morphometry, particularly for attention and working memory. These findings demonstrate that normative modelling captures meaningful individual differences and improves cognitive prediction after TBI.

Topic Area: Development, Individual Differences & Clinical Populations