By Marc Dubin, MD, PhD
Associate Professor of Clinical Psychiatry and Associate Professor of Research in Neuroscience, Weill Cornell Medicine
The authors showed that gamma frequency transcranial alternating current stimulation (tACS) targeting the precuneus in patients with Alzheimer’s disease (AD) improved measures of immediate and delayed verbal memory as well as associative memory for faces. Additionally, neurophysiological measurements of cholinergic transmission improved. Apolipoprotein E genotype and baseline cognitive performance were correlated with treatment response. Memory improvement was correlated both with increases in gamma frequency power in posterior cortical regions and spatial overlap between the modeled electric field distribution and the precuneus target site.
Benussi A, Cantoni V, Grassi M, et al. Increasing brain gamma activity improves episodic memory and restores cholinergic dysfunction in Alzheimer’s disease. Ann Neurol 2022;92:322-334.
Most research into the pathogenetic mechanism and symptomatic treatment for Alzheimer’s disease (AD) has focused on cholinergic and glutaminergic transmission. However, recent work has highlighted gamma band (30 Hz to 80 Hz) desynchronization as an early marker for AD.
Re-entrainment of gamma oscillations in animal models of AD has reversed amyloid burden and improved cognitive function. Given the limited efficacy of existing treatments, stimulating relevant cortical areas with gamma-band transcranial alternating current stimulation (tACS) is a promising treatment approach.
In this randomized, double-blind, crossover, sham-controlled study, Benussi and colleagues examined the effect of gamma-band tACS targeting the precuneus on symptoms of verbal learning and associative memory in patients with AD. Electroencephalogram (EEG) analysis was performed pre- and post-stimulation to determine if memory changes correlated with gamma-band entrainment. Finally, the following factors were considered as possible predictors of treatment response: electric field distribution relative to the precuneus stimulation site; apolipoprotein E (ApoE) genotype, a robust risk marker of late-onset AD; and brain-derived neurotrophic factor (BDNF) genotype, which has been associated with efficacy of transcranial stimulation.
Subjects had a diagnosis of AD confirmed by either cerebrospinal fluid analysis (Aβ1-42 ≤ 600 ng/L and tau ≥ 400 ng/L) or positive amyloid positron emission tomography scan. Individuals with history of cerebrovascular accident, traumatic brain injury, or serious medical illness were excluded. To ensure the safety of tACS, individuals with a history of seizures, metal implants in the head, or other electronic implants (e.g., pacemakers) were excluded. Subjects could be on a medication regimen, including cognitive enhancers, but doses had to be stable from at least six weeks prior to the stimulation protocol. tACS (active or sham) was applied to the precuneus region (Pz on a standard 10 to 20 EEG montage) at 40 Hz for 60 minutes. One week later, subjects were crossed over to the opposite group and received another 60-minute active or sham treatment.
Primary outcome measures were change in episodic memory (Rey Auditory Verbal Learning test [RAVL]) and associative memory (Face-Name Associations Task [FNAT]) from pre- to post-tACS, comparing active and sham groups and change in baseline EEG gamma activity from pre- to post-tACS. Secondary outcome measures were changes in baseline cholinergic transmission measured with a transcranial magnetic stimulation (TMS) paired pulse protocol producing motor-evoked potentials in the contralateral median nerve and correlation between changes in memory performance and apoE and BDNF genotypes.
Sixty subjects, average age 72.3 years, mean Mini-Mental State Examination (MMSE) score 23.9, and 51.7% female, received a full neuropsychological battery, structural magnetic resonance imaging (MRI) scan, and ApoE and BDNF genotyping. Subjects were randomized to active- and sham-tACS groups. All subjects completed the full protocol. The sham was effective, evidenced by the fact that equal numbers of active- and sham-receiving subjects reported tingling cutaneous sensations and no subject reported seeing visual phosphenes, which could occur from collateral visual cortex stimulation.
In the active-tACS group, RAVL immediate and delayed recall improved from pre- to post-tACS, while there was no effect of sham. Additionally, the active-tACS group improved in FNAT associative learning, administered for the final 20 minutes of the stimulation session, while the sham group did not. Restoration of cholinergic inhibition occurred with active-tACS, but not sham-tACS, and was correlated with improvements in RAVL delayed recall scores and FNAT scores.
These effects were specific both to the precuneus stimulation site as well as the cognitive domain of memory. A supplemental study targeting tACS to the right dorsolateral prefrontal cortex produced no significant effects on RAVL immediate or delayed scores or on FNAT scores. Similarly, a second supplemental study found no effects on executive function, verbal fluency, or visuospatial ability with precuneus stimulation.
The authors considered several possible predictors of memory improvement from active-tACS, including age, sex, education, cognitive reserve, MMSE, basic activities of daily living, neuropsychiatric inventory score, ApoE, and BDNF genotype. Of these, only ApoE genotype and MMSE score were significant. In both cases, markers of lower risk and milder disease predicted better improvement in memory. ApoE ε4 non-carriers had the largest improvement in memory scores while ApeE ε4 heterozygotes and ε4 homozygotes had progressively smaller improvements. Similarly, higher baseline MMSE score predicted larger improvement in memory scores. Active-tACS was associated with increased EEG power in the beta (12 Hz to 20 Hz) and gamma (20 Hz to 40 Hz) ranges in bilateral parietal leads, but not in lower frequency (theta and alpha) bands. Power increases were positively correlated with increases in RAVL delayed recall as well as increases in FNAT scores. Additionally, there was a dose effect of electric field strength. The strength of the electric field over the precuneus, determined by computational modeling in a subset of 13 subjects, correlated with improvement in RAVL immediate and delayed recall scores.
COMMENTARY
For learning and memory symptoms of AD, tACS using the gamma frequency band is a promising treatment. A single 60-minute stimulation session led to improvement of both episodic and associative memory. The case for tACS is made even more compelling by the fact that memory improvements were associated with gamma-band power increases in a parietal network strongly connected to the precuneus stimulation site, a dose effect of the field strength of electrical stimulation over the precuneus, and improvement in cholinergic neurotransmission. tACS appears to be most effective in individuals with the least ApoE genetic risk and at the earliest stage of cognitive decline, suggesting that it may be possible to direct this treatment to these individuals as part of a personalized AD treatment algorithm.
The meaningful effects on memory following such a short stimulation period of 60 minutes beg the question of whether a series of stimulation sessions would have a stronger and more durable effect on episodic and associative memory in AD patients.
