Optimizing tACS for working memory: differential outcomes in healthy aging and non-amnestic mild cognitive impairment.

Transcranial alternating current stimulation (tACS) in the theta frequency range has been shown to enhance working memory (WM) performance. However, no studies have directly compared the effects of theta tACS between cognitively healthy elderly (HE) subjects and subjects with non-amnestic mild cognitive impairment (MCI). Our proof-of-concept study investigated the effects of monofocal (frontal) and bifocal (frontoparietal) theta tACS on WM in two cognitive loads in HE subjects and in subjects with MCI. In this sham-controlled, single-blinded, repeated-measures study with counterbalanced stimulation order across subjects (n = 55), theta tACS (4.51 Hz, 1.5 mA, 20 min) was applied either over the frontal site alone or simultaneously over frontal and parietal sites. WM 2-back and 3-back tasks were performed during and after each stimulation session. In the HE group, both frontal and frontoparietal stimulations improved performance, albeit with load-dependent differences. Frontal stimulation was particularly effective in the higher cognitive load, enhancing accuracy (β = -3.87; p = .033) and reaction times (β = - .042; p = .002) in the 3-back task. Frontoparietal stimulation improved accuracy (β = -3.74; p = .027) but not reaction time (p > .22) in the 2-back task in the HE group. Frontoparietal stimulation enhanced accuracy in the 3-back task across all participants (β = 1.91; p = .043). In the MCI group, frontal stimulation led to faster reaction times in the 3-back task, although the effects were not robust. Lastly, a marginally significant improvement in reaction times was observed in a letter 2-back transfer task following frontal stimulation (β = - .034; p < .092) across all participants. Our findings indicate that theta tACS over the frontal and frontoparietal areas elicits benefits in WM performance, driven mainly by enhancements in HE subjects. The effects of stimulation varied with cognitive load and montage, suggesting that optimal stimulation parameters may differ depending on task demands. The non-amnestic MCI group did not exhibit greater improvements despite their lower baseline performance, possibly due to higher variability in pathology and compensation. Multiple sessions or alternative stimulation parameters may be needed to achieve robust effects in subjects with MCI. The study was retrospectively registered on ClinicalTrials.gov (NCT06563453).