Towards deep brain monitoring with superficial EEG sensors plus neuromodulatory focused ultrasound

Electrical recordings of neuronal activity in the brain allow the monitoring and interpretation of specific cortical processes on a millisecond time scale. This technology has contributed to the fundamental understanding of brain function along with other brain imaging techniques such as fMRI, but more importantly, has also enabled direct brain machine interfaces (BMI) for e.g. the control of prosthetic devices by patients suffering from paralysis or ALS. However, reliable BMI devices, that can be employed for real time control require direct implantation of electrodes in the brain and are thus highly invasive. Among the non-invasive techniques to image cortical activity, EEG, where electrodes are placed on a subjects scalp to record brain activity and fMRI are the most commonly used. FMRI possess good spatial resolution, but its poor time resolution (~ 1 second),

the fact that it measures blood flow as a proxy for neuronal activity and non-existent portability limit its ability as a BMI device and also as a tool to directly study cortical process which can happen on very fast timescales (< 12 ms). EEG on the other hand possess the necessary time resolution ( < 1ms), but suffers from poor spatial resolution and high susceptibility to artifacts originating from non-cortical sources, such as muscle activity (EMG), eye movements (EOG) or cardiac signals.

Here, we propose to overcome the fundamental limitations of EEG, specifically its poor spatial resolution and its vulnerability to non-cortical sources of electrical noise, by combining it with spatially focused (< 1 mm^3) ultrasound. This will provide a unique, spatially specific signal coming from only the US modulated region of cortex. In our approach the ultrasound focus serves the same purpose as the magnetic gradient fields in fMRI, by providing a unique frequency signature to a specific volume of cortex, which can be picked up and demodulated by the EEG. Like the magnetic fields, the ultrasound focus can penetrate scalp and skull noninvasively and can be centered on any part of the cortex. By modulating neuronal signals at non-physiological frequencies with ultrasound, we overcome the two fundamental limitations of non-invasive EEG, (1) we remove the spatial ambiguity of the signal and will attain millimeter scale precise resolution and (2) we make the EEG signal robust vs. artifacts, since the modulation frequencies can be outside e.g. the EMG or any other noise source spectrum.

We propose to test our approach by providing positive proof of this concept in an animal model, i.e. mice, where we elicit a well-defined and controlled cortical response, which we modulate with the ultrasound signal. Our hypothesis predicts the presence of this response in the US frequency demodulated signal.

Principal Investigator(s)
Research Lab