Brain Waves

Blood pressure and EEG results were measured to record the participants' physiological responses. Blood pressure (systolic (mmHg); diastolic (mmHg); and pulse rate) was measured using a sphygmomanometer (Omron, HEM-7011, Omron, China), and EEG results were recorded using a NeuroSky MindWave-EEG headset (Beijing Oriental Creation Technology Co., Ltd., China). The MindWave-EEG headset, which records brainwaves from the Fp1 position (frontal lobe) above the eye [26 ], is divided into four parts: a headband, an ear clip, a sensor arm containing the EEG electrode, and a Bluetooth device. Two dry sensors are used to filter and detect the EEG signals. The sensor tip identifies electrical signals of the brain from the forehead. The sensor detects ambient noise generated by human muscles, electrical sockets, computers, light bulbs, and other electrical devices. The ear clip acts as a ground and reference, which allows the ThinkGear chip to filter out the electrical noise [27 (link)]. The instrument measures the raw signal, power spectrum (high alpha, high beta), mediation level, and attention level. The raw EEG data are detected at a rate of 512 Hz. Other measured values are obtained every second. Therefore, the raw EEG data are the main source of information of EEG signals using the MindWave MW001. The device has small microchips that preprocess data and transfer electrical signals directly to the computer via Bluetooth. The raw EEG data which include high alpha and high beta power units were collected at 1-minute intervals at each experimental site, and 15-minute averages were compared between the two conditions. The headset can detect brainwave signals in the form of meditation and attention scores. According to the EEG e-Sense Metric, attention and meditation data are scaled from 1 to 100 (40 to 60: natural state, 60 to 80: slightly high, 80 to 100: very high, 20 to 40: slightly low, and 0 to 20: very low) [28 (link), 29 (link)]. The SDM [30 ] and STAI [31 ] were used to study the participant's physiological responses in both environments.

The code/software described in the paper is freely available online at https://github.com/singerlabgt/BrainWAVE. The code is available as Extended Data 1.

After constructing BrainWAVE stimulators, testing was performed to determine whether the devices generate appropriate stimulus intensity, timing, and other signal properties. Light illuminance and audio volume were measured with a light meter and decibel meter, respectively, with the distance between the sensor and meter approximating the distance from the sensory to the subjects’ eyes and ears (Extended Data Table 3-1). For mouse studies, light intensity was set at ∼150 lux and sound intensity at 60–65 dB (Garza et al., 2020 (link); Martorell et al., 2019 (link)). For human studies, we adjusted stimulus intensity for each subject based on tolerance, with the levels ranging from 0 to 1400 lux for brightness and 0–80 dBA for sound (He et al., 2021 (link)). We measured the frequency and duty cycle of the audio and visual stimuli in real-time using an oscilloscope connected to the analog output ports of the light and decibel meters (Fig. 3A). Alternatively, the timing of the light and sound stimulus may be measured with a photodiode and a microphone connected to an oscilloscope, or the stimulus may be recorded on a laptop and analyzed on a computer. Audio and visual signals were measured simultaneously to compare their duty cycle, frequency, and phase timing.
To modulate neural activity, we generated sensory signals at specific frequencies depending on the experimental design. Visual γ flicker (40 Hz) was produced using a 5.17-V, 40-Hz square wave with a 50% duty cycle (Fig. 3B). The voltage must be greater than 4 V to operate the MOSFET. Auditory γ flicker was produced with a pure sinusoid tone signal that was modulated by a 40 Hz square wave with a 50% duty cycle for audiovisual stimulation, and a 4% duty cycle for audio-only stimulation (Fig. 3C). The pure tone used was adjusted to fall within the center of the hearing range of the species tested: 10 kHz for mice and 7 or 8 kHz for humans (Heffner and Heffner, 2007 ). We used a 4% duty cycle for audio-only stimulation to more closely match the timing of clicks in studies on auditory steady-state responses evoked with 40-Hz click trains (Galambos et al., 1981 (link); Stapells et al., 1984 (link); Osipova et al., 2006 (link); Ma et al., 2013 (link); Thuné et al., 2016 (link)). Other frequencies of sensory were generated in a similar manner typically with a 50% duty cycle. Randomized stimulation was used to compare periodic to aperiodic flicker stimulation and had varying duty cycles (from 33% to 99%). Audio and visual signals were typically synchronized with similar duty cycles, but offset signals or different duty cycles may be desired in some cases (Fig. 3D–G,I).