Finally: thanks to @datalad@fosstodon.org for helping me manage this complex project across institutions. Decentralized neuroscientific data management is here!

Based on these results, we propose that older adults may be stuck in a suboptimal ‘middle ground’ that neither affords stable task selectivity when uncertainty is low, nor flexible task sensitivity in dynamic or uncertain contexts.

Intriguingly, in older adults, sensory “excitability” signatures were *especially* muted in their response to uncertainty, highlighting a potential failure point that can be targeted in more detail in future work.

Indeed, older adults showed a dampening of uncertainty adjustments across behavioral, EEG, fMRI, and pupil signatures. Supporting a “maintenance” account of brain aging, older individuals with more young-like neural recruitment could more flexibly select task-relevant features.

It is plausible that joint declines of task set management (e.g., via an MD thalamus – cingulate cortex (ACC) network), perception, and decision systems constrain the dynamic range of uncertainty adjustments in the aging brain. We attempt to comprehensively compare them here.

Decisions in more uncertain contexts used less info about any single feature, yet more information for unprobed features could be decoded in visual cortex: a precision-sensitivity trade-off. The dynamic range of this decision trade-off was reduced in older adults.

In the MRI scanner and during EEG acquisition, 100 younger and older adults performed a challenging multi-attribute task, in which they had to discriminate up to four different features in a stimulus.

Uncertainty contexts require flexible control.

(How) does our ability to manage task uncertainty differ in the aging brain?

Multimodal EEG/fMRI evidence (w. @garrettneuro.bsky.social, U. Lindenberger and U. Mayr) is now out in Nature Communications.

www.nature.com/articles/s41...