Putting the Pieces Together: Mental Construction of Semantically Congruent and Incongruent Scenes in Dementia

Scene construction refers to the process by which humans generate richly detailed and spatially cohesive scenes in the mind's eye. The cognitive processes that underwrite this capacity remain unclear, particularly when the envisaged scene calls for the integration of various types of contextual information. Here, we explored social and non-social forms of scene construction in Alzheimer's disease (AD; n = 11) and the behavioural variant of frontotemporal dementia (bvFTD; n = 15) relative to healthy older control participants (n = 16) using a novel adaptation of the scene construction task. Participants mentally constructed detailed scenes in response to scene-object cues that varied in terms of their sociality (social; non-social) and congruence (congruent; incongruent). A significant group x sociality x congruence interaction was found whereby performance on the incongruent social scene condition was significantly disrupted in both patient groups relative to controls. Moreover, bvFTD patients produced significantly less contextual detail in social relative to non-social incongruent scenes. Construction of social and non-social incongruent scenes in the patient groups combined was significantly associated with independent measures of semantic processing and visuospatial memory. Our findings demonstrate the influence of schema-incongruency on scene construction performance and reinforce the importance of episodic-semantic interactions during novel event construction.