Effects of 40 Hz transcranial alternating current stimulation (tACS) on cognitive functions of patients with Alzheimer’s disease: a randomised, double-blind, sham-controlled clinical trial

Gamma frequency stimulation is found to alleviate memory deficits on animal models of Alzheimer’s disease (AD), potentially by activating neuroimmune signalling and removing A-beta plaques in the brain.1 2 A feasible, translational hypothesis is that gamma band brain stimulation (eg, transcranial alternating current stimulation (tACS)) would yield clinical benefits on cognition in patients with AD. Here we performed a 6-week gamma tACS (2 mA, 40 Hz over bilateral temporal lobes) on a total of 50 subjects to elucidate clinical efficiency and safety of gamma stimulation for patients with AD. The study included a 6-week tACS intervention phase (5 days on and 2 days off for weekends) and another 12-week efficacy/safety assessment phase without tACS intervention (figure 1A). Bilateral tACS over temporal lobes (located by 10–20 electroencephalogram system) was delivered through saline-soaked sponges (sized 4×4 cm2) using a stimulator (Transcranial, London, UK). The 40 Hz sinusoidal current was applied at 2 mA with a duration of 20 min. Mini-Mental State Examination (MMSE) and Alzheimer’s Disease Cognitive Component Assessment (ADAS-Cog) were used for cognition evaluation before (0 week), at the end of (6 weeks) and 12 weeks (end +12 weeks) after the end of the stimulation. Serum A-beta levels were measured with ELISA. Figure 1 Effects of tACS on cognitive functions and serum Aβ levels in patients with AD. (A) Consolidated Standards of Reporting Trials diagram of the primary phases of the clinical trial. (B,C) MMSE score (B) and ADAS-Cog total score (C) at the designated time points. There was no significant difference between tACS group and sham group at baseline for either MMSE (F=0.60, p=0.44) or ADAS-Cog total score (F=0.09, p=0.77). Repeated measures analysis of variance revealed a significant time effect (MMSE: F(2,48)=33.91, p<0.01; ADAS-Cog: F(2,48)=10.84, p<0.01) and an interaction effect (time×group) (MMSE: F(2,48)=5.92, p<0.01; ADAS-Cog: F(2,48)=6.43, p=0.002), but not a group …