Impact of aging on theta-gamma phase-amplitude coupling during learning: A multivariate analysis.

Younger adults learned the visual sequence significantly faster than older adults.

• EEG was recorded from 113 younger and 117 older healthy adults during a sequence learning paradigm
• The paradigm involved 6423 repetitions and 55,944 stimuli total
• Participants learned a fixed visual sequence over repeated observations, allowing tracking of mPAC throughout incremental learning
• Learning speed differences were statistically significant between age groups

Multivariate phase-amplitude coupling (mPAC) increased with learning in both younger and older adults.

• mPAC was computed using generalized eigendecomposition (GED), which avoids confounds from non-sinusoidal waveforms
• mPAC captured coupling across distributed brain regions rather than at individual electrodes
• The increase in mPAC tracked incremental learning progress across repeated sequence observations in both age groups
• PAC was measured between mid-frontal theta and occipital gamma oscillations

mPAC distinguished fast from slow learners within both younger and older adult groups.

• Fast learners showed higher mPAC compared to slow learners
• This differentiation was observed across both age groups
• The finding suggests mPAC reflects individual differences in memory formation capacity regardless of age
• PAC between mid-frontal theta and occipital gamma is described as a proposed marker for the parallel storage of multiple items in working memory

Older participants showed overall reduced mPAC compared to younger adults, suggesting compromised parallel storage in working memory.

• The age-related reduction in mPAC was observed despite both groups showing learning-related increases
• Reduced mPAC in older adults is interpreted as reflecting compromised parallel storage in working memory
• The study included 117 older healthy adults and 113 younger healthy adults
• This finding extends prior research that had mainly focused on young adults with only a few studies in aging populations

Stratification analysis revealed that mPAC effects persisted across performance groups matched for mid-frontal theta power, indicating theta amplitude alone does not explain the observed mPAC effects.

• Performance groups were matched for mid-frontal theta power to control for amplitude confounds
• mPAC differences remained significant after controlling for theta power
• This addresses a key limitation of univariate PAC methods, which are susceptible to spurious estimates due to EEG nonstationarities
• The result supports the interpretation that phase-amplitude coupling reflects a distinct neural mechanism beyond simple power changes

The study applied multivariate PAC (mPAC) using generalized eigendecomposition (GED) to overcome limitations of univariate PAC methods.

• Conventional univariate PAC methods are susceptible to spurious estimates due to EEG nonstationarities
• Prior studies typically assessed PAC at individual electrodes, potentially overlooking broader functional significance across distant brain regions
• GED-based mPAC avoids confounds from non-sinusoidal waveforms
• mPAC captures coupling across distributed brain regions simultaneously