The other crucial part of the story is who gets the stimulation, and when they get it. The effects of stimulation like tDCS vary depending on both characteristic traits of the individual receiving stimulation, and also on their current state at the time of stimulation [ix]. This is demonstrated nicely by the finding that adults with maths anxiety are better at performing calculations during tDCS, while those without maths anxiety are impaired by tDCS [x].
However, we have to acknowledge that if stimulation can have an effect on the brain, then without a comprehensive understanding of it, we could have a dangerous tool on our hands. One reason stimulation could be dangerous is that brain regions typically do more than one thing. To give one example, the dorsolateral prefrontal cortex is an area of the brain approximately underlying each side of the hairline.
It is involved in high level planning and reasoning, and is often targeted in tES research including the study on number line understanding in children.
However, stimulation of this area has also been found to alter the likelihood of making high-risk, high-gain decisions [xi]. Another crucial consideration is the long-term effects of brain stimulation, especially in young children. This has not yet been addressed by the research. It seems pretty likely that transcranial brain stimulation can and does have an effect on cognition and behaviour.
However, there are just too many unknowns here to make predictions about classroom practice so, on balance, for the moment the verdict is neuro-miss. It may be more pertinent to talk about clinical practice, though.
For a highly insightful and entertaining first-hand account of tES, see this New Yorker article. Increasing human brain excitability by transcranial high-frequency random noise stimulation. Journal of Neuroscience, 28 52 , , doi: Transcranial random noise stimulation and cognitive training to improve learning and cognition of the atypically developing brain: A pilot study. Scientific Reports, 7 : Doi: Neurophysiological mechanisms of transcranial direct-current stimulation-enhanced motor learning in healthy children.
Clinical Neurophysiology, 3 , e Quantitative review finds no evidence of cognitive effects in healthy populations from single-session transcranial direct current stimulation tDCS. As such, phase-locked tACS or rTMS, although potentially useful in theory for enhancing stimulus potency and consistency, is difficult to accomplish in practice due to challenges in recovering the source signal for closed-loop stimulation.
Frontiers | A Real-Time Phase-Locking System for Non-invasive Brain Stimulation | Neuroscience
Here, we describe a novel apparatus for a closed-loop stimulation system that can provide real-time, phase-locked brain stimulation and can be applied with a wide range of neuromodulation techniques. The approach relies on the use of brief pulses of stimulation as are employed in rTMS, or in this case with brief-pulse tES , such that only a small proportion of the data samples in each cycle are contaminated by stimulus artifact c.
In this work, we have tested our method using pulsed-transcranial electrical stimulation ptES Alon et al. The intended scope of this paper is to describe the technique and provide a preliminary proof-of-concept in vivo demonstration. Subsequent work will examine in detail the behavioral effects of phase-locked vs. We have developed a closed-loop brain stimulation technique that is able to read electroencephalography EEG data, analyze it in real-time to extract specific phase and frequency information, and control the stimulator based on the phase and frequency of the underlying signal.
In this section we describe all components of this system in detail. The proposed system, consisting of EEG amplifier blue records the EEG signal and transmit the digital data to the computer. The computer green receives the EEG recordings and analyze the signal to extract the timing of the stimulation and communicates the instruction through a digital to analog converter to the stimulator.
The stimulator pink , provides the stimulation based on the information received and allows for closed loop brain stimulation. OpenVibe Renard et al.
The EEG system sampled the data at 2 kHz and downsampled the data to Hz in hardware, while applying the appropriate anti-aliasing filters. The data was buffered and sent 4 samples at a time to the OpenVibe software. The V-AMP amplifier used a bit data resolution with a wide dynamic range. This ensured that the amplifier was not saturated during the recording due to the stimulation and resulting artifacts.
The script was developed to analyze the EEG signal recorded from the EEG channel of interest Fz for theta recording and Oz for alpha recording and control the stimulator based on the method presented in our previous work Mansouri et al. In summary, this script first removes the stimulation artifact from the recorded EEG; second, it analyzes the EEG to extract the timing for the next stimulation pulse and, finally, it communicates with an Arduino Due board to control the stimulator output in real-time.
Accurate recovery and analysis of EEG activity occurring simultaneously with a large stimulation artifact is a challenging objective. For stimulation methods such as transcranial alternating current stimulation tACS , which delivers a continuous sine-wave stimulation pulse, closed-loop stimulation is particularly challenging, as it is difficult to recover the underlying EEG signal accurately, or determine its phase in a given frequency component during active stimulation.
An alternative, workaround approach is to minimize the proportion of EEG samples that are affected by stimulation artifact, by using a tES waveform consisting of short, square-wave pulses rather than continuous sinusoidal stimulation. During brief-pulse stimulation, only a small portion of the EEG samples are contaminated with the stimulation artifact, so that much of the EEG signal and in particular, its phase in a given frequency component remains recoverable over a given window of time.
For this reason, we used short pulsed stimulation in this work, which enabled us to assess the feasibility of closed-loop, phase-locked tES without the confounding presence of a stimulation artifact c. Neuropace US patent B2 , using readily available and inexpensive components commonly employed in laboratory settings.
The amplitude of the signal compared with its local median window size of 20 samples was used to detect the large artifact McNames et al.
Next, the signal contaminated with the artifact was deleted and replaced with an interpolated signal using a spline interpolation method Waddell et al. This method worked well only when the duration of the artifact was much smaller than the period of the underlying brain oscillation. For the purposes of this proof-of-concept study, we selected theta 4—8 Hz and alpha 8—13 Hz bands for testing this closed-loop phase locking method. To assess the artifact removal method performance, stimulation at arbitrary phase and frequency was applied to study the artifact duration and the artifact detection method.
Pulses of 5 ms duration were delivered at random intervals of — ms. Averaging of epochs of stimulation artifact recordings, time-locked to the onset of the stimulation pulse, was used to assess the lasting effect of the artifact after the onset of the pulse see Results below. Artifacts from pulses were averaged to study the true effect of the stimulation on the recording. A The epochs and their average. B An example of the pulse detector, asterisk showing location of the detected pulses.
C,D Pulses removed and signal interpolated using spline interpolation. As described in our previous work Mansouri et al. Based on our previous simulations offline, we used a 10th order elliptical filter, converted the quantized filter to second-order sections, and applied it to the recording. The filtered signal was first zero-padded to increase the bin resolution in the FFT.
The FFT bin with maximum power in the frequency range of interest was selected as the dominant frequency and its phase and frequency were used to calculate the timing of the next stimulation pulse. To correct for the small delays in the system or a small phase shift that was introduced by the filter or other components of the signal processing or hardware delays, an empirically calculated correction time was added to the calculated pulse time.
In this system, the stimulator follows the Arduino-generated voltage waveforms and produces a current-controlled output proportional to that voltage. In order to demonstrate the potential applicability of the present technique across different brain regions, different states resting vs. Occipital regions are particularly potent and well-studied generators of resting alpha rhythms, particularly during the resting eyes-closed state, and are thus widely used in studies of EEG alpha activity Lehmann, ; Vossen et al. Likewise, midline frontal regions are particularly potent and well-studied generators of theta rhythms, particularly during cognitive control tasks, which are likewise widely used in EEG studies of cognition in healthy controls and individuals with illness Frank et al.
Such studies commonly employ evoked potentials rather than resting-state activity Frank et al. The system was tested by providing phase-locked stimulation at theta 4—8 Hz and alpha 8—13 Hz in mid-frontal and occipital human brain regions, respectively. The testing was performed during eyes closed EEG for alpha-band testing 5 min.
For theta-band testing sessions 5 min , the participants played a computer-based reinforcement learning game Frank et al.
In this reinforcement learning game, participants were presented with pairs of Japanese characters and asked to choose one by pressing left or right key on a keyboard, followed by a visual feedback won or lost Frank et al. Each stimulation alpha and theta was applied for 50 pulses for each of 0 o , 90 o , o and o phase angles. Phase-locking values for each band and phase-angle were computed as below, along with their distributions for each phase and band.
Participants were 5 healthy volunteers 3 male, 2 female, ages 27—30, mean age To measure the phase of the underlying brain oscillation during the stimulation, the stimulation artifact was first removed and the signal interpolated using the methods described earlier. Then, the bandpass filter was applied. The phase of the EEG was measured as the angle of the Hilbert transformation during the stimulation pulse. We use polar histogram plots to visualize the phase of the stimulation compared to the phase of the underlying brain oscillations.
As noted above, stimulation at arbitrary phase and frequency was applied to study the artifact duration and the artifact detection method. Pulses of 5ms duration were delivered at random intervals of — ms.
Averaging epochs of stimulation artifact recordings, time-locked to the onset of the stimulation pulse, illustrates the lasting effect of the artifact after the onset of the pulse Figure 2A. Thus, our artifact removal method is set to remove 20 ms of the recording and interpolate the signal for that duration Figures 2B—D. Alpha and theta stimulation were conducted on a healthy volunteer to test the system delays and calculate the angle corrections for each of the stimulations.