Feasibility of a Within-Subject fMRI Design for Eyeblink Conditioning in Future Treatment Studies

Background. Delay eyeblink conditioning is a well-established paradigm for evaluating cerebellum-dependent associative motor learning and may be useful for assessing treatments targeting the cerebellum. This preparatory study examined the feasibility of a within-subject eyeblink conditioning design during fMRI, asking whether conditioning could be performed twice within the same participants with comparable behavioral outcomes and activation patterns. Whole-brain fMRI allowed us to re-evaluate the cerebello-cerebral network involved. Methods. Twenty healthy participants completed two delay eyeblink conditioning sessions in a 3T MRI scanner. An air-puff served as the unconditioned stimulus, and auditory and visual conditioned stimuli were counterbalanced between sessions. Conditioned response incidences (CRIs) were quantified using pupillometry. Event-based fMRI analysis was based on unreinforced CS-only trials during acquisition. Test-retest reliability was assessed for behavioral and neural measures. Results. CRIs increased across acquisition blocks in both sessions, with no significant difference between auditory and visual CSs. CRIs were higher in Session 2 and showed moderate test-retest reliability (ICC = 0.523). Early Session 1 CRIs allowed classification into Fast and Slow learners . Eyeblink conditioning-related activations involved a broad cerebellar, subcortical, and cortical network consistent with the action-mode network and primary motor regions. Neural test-retest reliability was moderate to good in several regions (local ICC maxima 0.50–0.87). Conclusions. Pupillometry enabled assessment of conditioned eyeblink responses during fMRI. Learning-related behavioral and neural effects were observed in both sessions, supporting feasibility for future within-subject treatment studies. However, learning transfer and session effects need to be considered.