Cerebral glucose metabolic correlates of cognitive and behavioural impairments in amyotrophic lateral sclerosis

Objective Half of ALS patients are cognitively and/or behaviourally impaired.

As cognition/behaviour and cerebral glucose metabolism can be correlated by means of ^18F-Fluorodeoxyglucose positron emission tomography (FDG-PET), we aimed to utilise FDG-PET, first, to replicate group-level differences in glucose metabolism between non-demented ALS patients separated into non-impaired (ALSni), cognitively impaired (ALSci), behaviourally impaired (ALSbi), and cognitively and behaviourally impaired (ALScbi) groups; second, to investigate glucose metabolism and performance in various cognitive domains; and third, to examine the impact of partial volume effects correction (PVEC) of the FDG-PET data on the results.

Methods We analysed neuropsychological, clinical, and imaging data from 67 ALS patients (30 ALSni, 21 ALSci, 5 ALSbi, and 11 ALScbi).

Cognition was assessed with the Edinburgh Cognitive and Behavioural ALS Screen, and two social cognition tests.

FDG-PET and structural MRI scans were acquired for each patient.

Voxel-based statistical analyses were undertaken on grey matter volume (GMV) and non-corrected vs. PVE-corrected FDG-PET scans.

Results ALSci and ALScbi had lower cognitive scores than ALSni.

In contrast to both ALSni and ALSci, ALScbi showed widespread hypometabolism in the superior- and middle-frontal gyri in addition to the right temporal pole.

Correlations were observed between the GMV, the FDG-PET signal, and various cognitive scores.

The FDG-PET results were largely unaffected by PVEC.

Interpretation Our study identified widespread differences in hypometabolism in the ALScbi-ni but not in the ALSci-ni group comparison, raising the possibility that cerebral metabolism may be more closely related to the presence of behavioural changes than to mild cognitive deficits.