Beyond episodic memory: Semantic processing as independent predictor of hippocampal/perirhinal volume in aging and mild cognitive impairment due to Alzheimer's disease

Objective: Given that lexical-semantic decline precedes episodic memory deficits in the Alzheimer's disease (AD) timeline, it is expected that performance on a lexical-semantic task would be associated with mediotemporal volumes independently of the association this region has with episodic memory in the early stage of AD. Method: Fifty patients with mild cognitive impairment due to AD and 50 healthy adults completed tests of lexical-semantic skills (category fluency test), episodic memory for semantically relevant material (prose memory test), episodic memory for non semantically relevant material (Rey-Osterrieth Figure test), lexical-executive abilities (letter fluency test), and a neurostructural MRI. Hippocampal, perirhinal, entorhinal, temporopolar, and orbitofrontal volumes were extracted. The association between test performance and volume of each region was tested using partial correlations (age-education corrected). The improvement (-R2) in predicting volumetric indices offered by episodicmemory/ lexical-semantic processing, once accounting for their counterpart, was tested using hierarchical regressions. Results: There were no significant findings for control indices. Prose memory accounted for independent portions of volumetric variability within almost all regions. Category fluency accounted for independent portions of volumetric variability of left and right hippocampus and left perirhinal cortex in addition to the predictive strength of the Rey-Osterrieth Figure, and for an independent portion of volumetric variability in the left hippocampus in addition to the predictive strength of prose memory. Conclusions: There was an association between hippocampal and perirhinal volume and lexical-semantic processing in addition to the contribution given by episodic memory. This statistical separation supports the importance of lexical-semantic processing as independent indicator of AD.