Phase-dependent closed-loop modulation of neural oscillations in vivo

Abstract

Normal brain function is associated with an assortment of oscillations of various frequencies, each reflecting the timing of separate computational processes and levels of synchronization within and between brain areas. Stimulation accurately delivered on a specified phase of a given oscillation provides the opportunity to target individual aspects of brain function. To achieve this, we have developed a highly responsive system to produce a continuous online phase-estimate. In addition to stable oscillations, the system accurately tracks the early cycles of short, transient oscillations and can operate across the frequency range of most established neuronal oscillations (4 to 250 Hz). Here we demonstrate bidirectional modulation of the pathologically elevated parkinsonian beta-band oscillation (around 35 Hz) in 6-OHDA hemi-lesioned rats. Beta phase, monitored using a single channel electrocorticogram above secondary motor cortex, was used to drive electrical stimulation of the globus pallidus on one of eight phases spanning the oscillation cycle. Stimulation of the early ascending phase suppressed the oscillation whereas stimulation of the early descending phase was amplifying. By implementing a rule that prevented stimulation when the phase estimate was unstable, we achieved a system that could adapt stimulation rate and pattern to respond to the changes produced in the target oscillation. This allowed the electronic system to create and maintain a state of equilibrium with the biological system resulting in continuous stable modulation of the target oscillation over time. These results demonstrate the feasibility of phase locked stimulation as a more refined strategy for remediation of pathological beta oscillations in the treatment of the motor symptoms of Parkinson's disease. Furthermore, they establish the utility of our algorithm and allow for the potential to assess the contribution of rhythmic activity in neuronal computation across a number of brain systems.

Parkinson's disease uncovers an underlying sensitivity of subthalamic nucleus neurons to beta-frequency cortical input in vivo

“…For other approaches, it could be possible to utilise the lower amplitude information. Several authors have postulated that phase-dependent stimulation could be an effective therapeutic strategy for treating motor symptoms ( Rosin et al, 2011 ; Azodi-Avval and Gharabaghi, 2015 ; Holt et al, 2016 ; Meidahl et al, 2017 ) and we have recently demonstrated the potential efficacy of this approach ( Holt et al, 2019 ; McNamara et al, 2020 ).…”

Real-time estimation of phase and amplitude with application to neural data

“…Additionally tracking phase may augment therapeutic benefit, though, by delivering stimulation pulses at specific points within the oscillatory cycle. This has been shown to effectively disrupt pathological oscillations [10][11][12]17 . So far, to our knowledge, adaptive stimulation algorithms integrating both the phase and amplitude of a signal have not been established yet.…”

“…A special application is developing efficient sensing algorithms for adaptive deep brain stimulation, a recent advancement of a widely used treatment option for Parkinson's disease and other neurological disorders 5 . One of the directions in this development is to adjust stimulation parameters according to a peripheral or neurophysiological signal's phase and amplitude computed on the fly [6][7][8][9][10][11][12] . The popular approach to phase and amplitude estimation is to exploit the analytic signal approach based on the Hilbert Transform (HT) or, equivalently, the wavelet transform with a complex wavelet 1,[13][14][15] .…”

Closed-loop optogenetic control of normal and pathological network dynamics

“…Similarly, adaptive DBS for Parkinson's disease is being trialled in humans, but this is currently limited to adjusting the amplitude of open-loop stimulation (Arlotti et al, 2018). However, phase-dependent electrical stimulation delivered to the globus pallidus has recently been used to disrupt pathological oscillations recorded in the overlying cortex of rodents (McNamara et al, 2020), while optogenetics would allow such stimulation to be controlled in realtime by abnormal activity recorded from the local network without interference from artefacts. Optogenetic therapies for epilepsy have been demonstrated in rodent models, but thus far these consist of continuous inhibitory stimulation delivered once a seizure is detected (Paz et al, 2013, Krook-Magnuson et al, 2013.…”