Raw number of face-name pairs correctly recalled with a maximum of 15 points; higher values are better at each time point. Change at post-session 5 (day 5 after baseline)) calculated relative to baseline performance (positive differences indicate improvement; negative values indicate decline). Data coding errors were identified and accurate values were reported 2/6/2023.

Patient-level factors of age, sex, race, RBANS Delayed Memory Index, MoCA, Emory version of the Wisconsin Card Sorting Test were included as covariates.

Performance measured using number of correctly identified locations (3 locations per stimulus; 15 total stimuli). Higher values indicate better performance. Change at post-session 5 (day 5 after baseline) calculated relative to baseline performance (positive differences indicate improvement; negative values indicate decline). Data coding errors were identified and accurate data were updated 2/6/2023.

Patient-level factors of age, sex, race, RBANS Delayed Memory Index, MoCA, Emory version of the Wisconsin Card Sorting Test were included as covariates.

Changes in task related blood oxygen dependent signal (BOLD) activation for the face-name (novel post > novel pre) contrast in the left inferior frontal gyrus (pars triangularis, pars orbitalis, pars opercularis). Data are preliminary betaweights for the above noted contrast. Positive values reflect increased BOLD signal while negative values represent reduced BOLD signal. Not all participants were able to complete fMRI, which explains sample size discrepancies with other outcome measures. Data coding errors were discovered and accurate, updated data reported on 2/6/2023

Performance on the Ecological Memory Simulations- Medical Instructions subtest. Raw points where higher values reflect better performance at each time point (0-15 possible points at each time point). Reported values reflect change from baseline (i.e., post-session day 5 vs. baseline) where positive values represent improvement and negative values represent decline. Corrected values are now included (2/23) that account for age, sex, and baseline neuropsychological abilities.

Changes on the Multifactorial Memory Questionnaire - strategy subscale. Raw points where higher values reflect better performance at each time point (0-76 possible points at each time point). Reported values reflect change from baseline (i.e., post-session day 5 vs. baseline) where positive values represent improvement and negative values represent decline.

Performance on Ecological Memory Simulations routes subtest (serial order). Higher values indicate better performance at each time point (0-9 possible points at each time point). Change from baseline is reported (post-session day 5 vs. baseline) so higher values indicate better recall while negative values indicate decline. Corrected data are now reported (2/23) accounting for age, sex, and baseline neuropsychological abilities.

Planned analyses to examine patient specific characteristics that affect treatment efficacy and would be vital for clinical translation at the individual patient level. Data analyzed using a composite "recognition" score that combined the face-name and object-location task recognition scores (total on a 0-30 point scale; values below are change from baseline; higher values are "better" and reflect improved memory test performance). Patient-level factors of age, sex, race, RBANS Delayed Memory Index, MoCA, Emory version of the Wisconsin Card Sorting Test (total sorts) were included as covariates.

linear mixed model Data analyzed using a composite "recognition" total score that combined the face-name and object-location task (change on 0-30 point scale relative to baseline) higher scores are better). Patient-level factors of age, sex, race, RBANS Delayed Memory Index, MoCA, Emory version of the Wisconsin Card Sorting Test (total sorts) were included as covariates.

Outcome measure is the change in betaweight (as a percent signal change: post session 5 minus baseline) during the face-name and object-location tasks (same metric as Primary Outcome measure 3) while controlling for the electric field (EF). Finite element model based measurement of electric field in the targeted brain regions (Values range from 0 to no theoretical upper limit with higher values reflecting more electrical current; most values will be under 0.5 V/m). The EF was calculated using the baseline MRI T1 scan for each individual. EF values were then included in the linear mixed model analysis of fMRI Beta-weight change (post training vs. baseline) since EF values at the targeted brain region varied across participants. Note, EF values only apply to active HD-tDCS groups, so sham groups have no data to report and were removed accordingly (updated 2/23).

linear mixed model that includes EF as a predictor variable as well as covariates of age, sex, education, and race.