A potential treatment for Alzheimer’s disease relies on 40 Hz light or sound to entrain gamma rhythms in the brain. This intervention, pioneered by Li-Huei Tsai and colleagues at Massachusetts Institute of Technology, Boston, and repeated by other research groups, was reported to lower amyloid and improve memory in mouse models. Now, in the March 6 Nature Neuroscience, researchers led by György Buzsáki at New York University’s Langone Medical Center question the robustness of this research. Buzsáki and colleagues found that 40 Hz white light had but a small effect on the gamma rhythms of neurons in the visual cortices of mouse models of amyloidosis. The treatment nudged down Aβ42 production, but did not budge plaque. “Our findings suggest that entrainment of natural gamma-band oscillations is not a likely mechanism for reducing AD pathology,” the authors wrote.
- Disputing prior research, 40 Hz light left plaques untouched in two mouse lines.
- Nor did it produce robust gamma rhythms in visual cortex.
- Some scientists wonder why a simple intervention has not been more widely adopted.
“Findings like these are an important reminder of why we need to have replication studies,” Bryce Mander at the University of California, Irvine, told Alzforum. Others agreed the new data raise questions about how sensory entrainment affects the brain, and under what circumstances it may work. Several scientists noted that methodological differences between the original report and the new research could explain the discrepancies, and said more studies are needed to parse these out. “This paper will reignite exciting debates on the validity and mechanisms of 40 Hz frequency light flicker treatments,” predicted Sheng-Tao Hou at the Southern University of Science and Technology in Shenzhen, China (comments below).
Tsai’s lab sparked this field of research with an initial report that flashing light at mice for an hour cut soluble Aβ42 in their visual cortices by half, activated microglia to attack amyloid and, with a week of daily treatment, mopped up two-thirds of plaques in said cortex (Dec 2016 news). Adding 40 Hz sound to the light was subsequently reported to broaden the effects to other brain areas and strengthen memory (Mar 2019 news).
Buzsáki and colleagues, who include scientists in the labs of Martin Sadowski and Wen-Biao Gan at Langone, set out to replicate the original 2016 study. A systems neuroscientist, Buzsáki was the first to identify the mechanisms underlying gamma and theta oscillations; he wrote an oft-cited book on brain rhythms and how they contribute to cognition.
For the present study, first author Marisol Soula exposed 14 APP/PS1 and eight 5XFAD mice to an hour of flickering light, then evaluated amyloid in their visual cortices via immunohistochemistry. She found no significant difference compared with plaque area in 11 APP/PS1 and nine 5XFAD controls. These were the same mouse models used in Tsai’s original study by Iaccarino et al., though they measured soluble Aβ40 and Aβ42 via ELISA, rather than plaque, after one hour of light exposure.
In a more direct comparison with the previous work, Soula and colleagues exposed a dozen 7-month-old 5XFAD mice to daily 40 Hz light for one week, then measured both Aβ and plaque in the visual cortex. The authors found a slight drop in both; it did not reach statistical significance overall, though the change in Aβ42 in the male mice did. In addition, Soula et al. crossed 5XFAD mice with transgenics that had fluorescent microglia, and examined the visual cortex using two-photon microscopy before and after light stimulation in six mice. Unlike in the original study, microglial morphology did not change.
Commentators offered varied explanations for the discrepancies. Annabelle Singer at Georgia Tech and Emory University, Atlanta, who was a co-first author on the 2016 paper from Tsai’s lab, noted that Buzsáki and colleagues used fixed rather than fresh tissue for their ELISAs to quantify Aβ, potentially lowering the sensitivity. Brendan Lucey at Washington University in St. Louis, Missouri, suggested that the larger number of mice used by Buzsáki's group may lend the new findings more weight—Iaccarino had used eight mice per group in the one-week paradigm. Brian Bacskai at Massachusetts General Hospital, Charlestown, praised the rigor of Buzsáki’s paper. “Negative results … always beg the question of whether the exact methods were followed. The expertise of the team, however, suggests that their results should be not be dismissed,” he wrote (comments below).
In addition to examining the effect of flashing light on amyloid, Soula and colleagues also measured its impact on electrical activity. By inserting probes into the brains of living mice and recording from them, they found that light exposure caused more than one-quarter of neurons in the visual cortex to oscillate in unison. Most of these were inhibitory interneurons. The stimulus had little effect on deeper brain regions, entraining only 7 percent of hippocampal neurons.
For context, Buzsáki noted that simply placing mice into a new environment causes a much greater response in the hippocampus, with many more neurons synching into physiological gamma rhythms. “By the logic of the Iaccarino et al. study, one expects that Aβ should be reduced even more by exploration than by being exposed passively to 40 Hz flashes,” he wrote to Alzforum. Intriguingly, slower 4 Hz theta waves are easier to entrain. Analyzing a previous data set (Steinmetz et al., 2019), Soula found that visual stimulation at 4 Hz recruited almost half of visual cortex neurons, and up to a third of hippocampal neurons.
“The most important lack of replication to me is the failure to robustly entrain brain rhythms … I do think this finding casts some doubt on gamma stimulation as a viable intervention,” Mander wrote to Alzforum. For her part, Tsai thinks this is not a major issue for the therapy. “We agree with the authors that native gamma oscillations, and the steady-state oscillations evoked by 40 Hz sensory stimulation, are likely distinct neurophysiological phenomena,” she wrote.
Buzsáki and colleagues suggested the lack of widespread entrainment might be due to the animals’ dislike of this flickering light. Other work indicates that sensory entrainment works best when the organism pays attention to a stimulus that has some behavioral relevance (Tiitinen et al., 1993; Williams et al., 2004). In Soula’s hand, the mice actively avoided the stimulus; when forced to experience it, cholinergic activity in the hippocampus spiked, a sign of behavioral stress. Enhancing cholinergic activity with acetylcholinesterase inhibitors is an approved AD treatment shown to improve cognition slightly.
Commenters thought the behavioral response could be a factor. “Aversion behavior may reduce gamma entrainment, and thus amyloid clearance,” Ki Woong Kim at Seoul National University, South Korea, wrote to Alzforum. Tsai noted that mice in her lab's experiments did not show signs of behavioral stress.
Tarek Rajji at the University of Toronto said the correlation between avoidance and poor outcomes could guide translational research. “This finding is consistent with the perspective from human studies that noninvasive brain stimulation offers best value when combined with behavioral interventions that motivate and engage participants, such as cognitive remediation,” he wrote. In other words, people not only have to see the light, but they have to want to see the light. Performing cognitive tasks during light and sound stimulation has been reported to extend gamma entrainment into the hippocampus (Khachatryan et al., 2022).
Researchers lauded Buzsáki and colleagues for their efforts to reproduce the earlier work. “Given the bias toward publishing only positive findings, I’m pleased to see a negative study investigating an important topic. This is how science should work—good investigators seek to replicate (or not replicate) each other’s findings. This generates conversation, debate, and collaboration to identify variables that might explain the discrepancy,” said Michael D. Fox at Brigham and Women’s Hospital, Boston.
Bacskai agreed. “This follows on a very high impact paper showing remarkable effects in mouse brain, and subsequent translation to human studies. I'm honestly surprised that we haven't seen 20+ papers investigating this approach, as it is so simple to reproduce by almost any lab with access to mice, and it would completely change how we treat patients with AD," he wrote.
“The best model of Alzheimer’s disease is a patient with Alzheimer’s disease. Different mouse models and mouse experiments can be helpful, but in the end it’s what happens in patients that is most important,” Fox said. Cognito Therapeutics, a biotech co-founded by Tsai and Ed Boyden at MIT, has been running several small human studies since 2018. In December 2022, the company began recruiting for a pivotal Phase 3 trial that aims to enroll 345 participants (Apr 2021 news; Sensory Stimulation Systems).—Madolyn Bowman Rogers
- Flashy Treatment Synchronizes Neurons, Lowers Aβ in Mice
- Flash! Beep! Gamma Waves Stimulate Microglia, Memory
- Does Synchronizing Brain Waves Bring Harmony?
Research Models Citations
- Steinmetz NA, Zatka-Haas P, Carandini M, Harris KD. Distributed coding of choice, action and engagement across the mouse brain. Nature. 2019 Dec;576(7786):266-273. Epub 2019 Nov 27 PubMed.
- Tiitinen H, Sinkkonen J, Reinikainen K, Alho K, Lavikainen J, Näätänen R. Selective attention enhances the auditory 40-Hz transient response in humans. Nature. 1993 Jul 1;364(6432):59-60. PubMed.
- Williams PE, Mechler F, Gordon J, Shapley R, Hawken MJ. Entrainment to video displays in primary visual cortex of macaque and humans. J Neurosci. 2004 Sep 22;24(38):8278-88. PubMed.
- Khachatryan E, Wittevrongel B, Reinartz M, Dauwe I, Carrette E, Meurs A, Van Roost D, Boon P, Van Hulle MM. Cognitive tasks propagate the neural entrainment in response to a visual 40 Hz stimulation in humans. Front Aging Neurosci. 2022;14:1010765. Epub 2022 Oct 6 PubMed.
- Soula M, Martín-Ávila A, Zhang Y, Dhingra A, Nitzan N, Sadowski MJ, Gan WB, Buzsáki G. Forty-hertz light stimulation does not entrain native gamma oscillations in Alzheimer's disease model mice. Nat Neurosci. 2023 Apr;26(4):570-578. Epub 2023 Mar 6 PubMed.