The CRUNCH model does not account for load-dependent changes in visuospatial working memory in older adults: Evidence for the file-drawer problem in neuroimaging

Abstract Numerous neuroimaging studies have shown that older adults tend to activate the brain to a greater extent than younger adults during the performance of a task. This is typically interpreted as evidence for cognitive compensation. The Compensation-Related Utilisation of Neural Circuits Hypothesis (CRUNCH) model is a highly influential model of compensation, and states that increased functional magnetic resonance imaging (fMRI) activity in older adults compared to younger adults should reverse at higher levels of task difficulty. We tested the CRUNCH model using a visuospatial working memory paradigm, and found that fMRI activity in older vs. younger adults was in the opposite direction to that predicted by the model. Given that the CRUNCH model is the predominant model of compensation, this result was surprising. We followed up our results with a systematic review of the CRUNCH in healthy ageing literature using p-curve analysis. We find evidence for selective reporting, or the ‘file-drawer’ problem, in the cognitive compensation literature. Further experimental work is required to validate the CRUNCH model in cognitive ageing. Abbreviations CRUNCH: compensation-related utilisation of neural circuits hypothesis; fMRI: functional magnetic resonance imaging Highlights - CRUNCH is the leading model of cognitive compensation in ageing - We find fMRI activity in old vs. young adults in opposite direction predicted by CRUNCH - We report quantitative evidence of selective reporting in CRUNCH literature