Exploring neural correlates of automated speech-based cognitive markers through resting-state functional connectivity in aging and at-risk Alzheimer's disease.

Greater language network connectivity in the left middle temporal gyrus was significantly associated with increased semantic verbal fluency temporal cluster switching after correction for multiple comparisons.

• This was the only finding to survive family-wise error (FWE) correction at p < .05
• Cluster size = 12 voxels, mean T = 3.86
• The association was found in the left middle temporal gyrus within the language network
• Temporal cluster switching refers to a speech feature derived from automated phone-administered semantic verbal fluency (SVF) tasks

No significant associations were found between cognitive composite scores and functional connectivity in the exploratory analyses.

• Exploratory analyses used an uncorrected threshold of p < .01
• Speech-based composite scores of cognition were derived from automated phone-administered SVF and verbal learning tasks (VLT)
• The lack of association held across the examined intrinsic connectivity networks
• Networks were identified via independent component analysis and dual regression

Individual SVF and VLT speech features showed exploratory, network-specific associations across executive, language, and default mode networks.

• These associations were examined at uncorrected p < .01 and did not survive correction for multiple comparisons
• The associations were described as 'spatially distinct connectivity patterns'
• Both semantic verbal fluency (SVF) and verbal learning task (VLT) features were involved
• Network-specific associations spanned executive, language, and default mode networks

The study analyzed data from 129 participants of the German PROSPECT-AD study, ranging from cognitively healthy individuals to those with mild cognitive impairment.

• The sample represented the Alzheimer's disease spectrum from cognitively healthy to mild cognitive impairment (MCI)
• Speech-based cognitive scores and features were derived from automated phone-administered tasks
• Tasks included semantic verbal fluency (SVF) and verbal learning tasks (VLT)
• Resting-state fMRI was used to assess functional connectivity (FC)
• Permutation-based voxel-wise regression was used, controlling for demographic and clinical covariates

Seed-to-voxel analyses were conducted to support network identification and complement the independent component analysis findings.

• Intrinsic connectivity networks were primarily identified via independent component analysis (ICA) and dual regression
• Seed-to-voxel analyses served as a complementary methodological approach
• This dual approach was used to strengthen confidence in network-level associations

The authors concluded that digital speech-based assessments have limited current utility for detecting functional connectivity alterations in at-risk individuals.

• Further validation using complementary methodological approaches was recommended
• Shorter intervals between fMRI and speech assessments were identified as a methodological need
• Testing in independent cohorts was cited as essential to establish reliability and clinical relevance
• The paper emphasized the need for validation 'for monitoring brain network changes'