Spatially specific, closed-loop infrared thalamocortical deep brain stimulation

Abstract

Deep brain stimulation (DBS) is a powerful tool for the treatment of circuitopathy-related neurological and psychiatric diseases and disorders such as Parkinson's disease and obsessive-compulsive disorder, as well as a critical research tool for perturbing neural circuits and exploring neuroprostheses. Electrically-mediated DBS, however, is limited by the spread of stimulus currents into tissue unrelated to disease course and treatment, potentially causing undesirable patient side effects. In this work, we utilize infrared neural stimulation (INS), an optical neuromodulation technique that uses near to mid-infrared light to drive graded excitatory and inhibitory responses in nerves and neurons, to facilitate an optical and spatially constrained DBS paradigm. INS has been shown to provide spatially constrained responses in cortical neurons and, unlike other optical techniques, does not require genetic modification of the neural target. We show that INS produces graded, biophysically relevant single-unit responses with robust information transfer in thalamocortical circuits. Importantly, we show that cortical spread of activation from thalamic INS produces more spatially constrained response profiles than conventional electrical stimulation. Owing to observed spatial precision of INS, we used deep reinforcement learning for closed-loop control of thalamocortical circuits, creating real-time representations of stimulus-response dynamics while driving cortical neurons to precise firing patterns. Our data suggest that INS can serve as a targeted and dynamic stimulation paradigm for both open and closed-loop DBS.

Keywords: Closed-Loop DBS; Deep Brain Stimulation; Deep Reinforcement Learning; Infrared Neural Stimulation.

Figure 1.

Implantation and EEG-MLR procedures. A. Left: Rodents were implanted with fiber optic optrodes into the medial geniculate body and 16 channel microwire arrays into auditory cortex. Placement of microwire array was confirmed by tonic single unit responses evoked from 80 dB filtered Gaussian noise stimuli during implantation. A. Right: Schematic of the 4-channel EEG-MLR recording preparation. B. Left: Schematic of the rodent auditory thalamocortical circuit. Stimulation optrodes were placed in the ventral division of the medial geniculate body with primary excitatory efferent projections to layer 3–4 of primary auditory cortex. Microwire array recording electrodes were placed in layer 3–4 of primary auditory cortex confirmed during surgery by low-latency single unit activity. B. Right: Histological images demonstrate placement of stimulation optrode was within medial geniculate body. C. EEG-MLR pre-post surgical ratios show small changes in wave P1, N1, and P2 correlates of auditory thalamocortical function in amplitude and latency due to passive presence of device at 65 or 85 dB-SPL click stimuli. While changes in amplitudes and latencies were observed effects, differences did not rise to level of significance (p>0.05). Rodent implantation and EEG diagrams were created using BioRender under publication license.

Figure 2.

A. Example INS-evoked peristimulus time histograms. BARS estimates of 1.8mJ (solid blue line) and 0.92mJ (dotted orange line) show higher energy pulses drive higher firing, lower latency responses. B. Bayesian hierarchical linear regression models of cortical dose-response profiles elicited from varying INS parameters. Distributions of regression parameters are given for applied laser energy, laser pulse width, and laser energy-pulse width interactions. Regressions show that increases in applied energy significantly increase maximum cortical firing rates with a maximum a priori estimate 0.58 increase in log firing rate in response to increases in log energy (95% HDI does not overlap 0). The width of the 95% HDI of the energy parameter (0.27–0.88) suggests that while cortical firing rates increase with increases in laser energy, INS dose-response profiles are dependent on the physiology of the neuron. Slight decreases in firing rate with increased laser pulse widths were observed (MAP = −0.055), but not significant (95% HDI overlaps 0). Laser energy and pulsewidth interactions also did not significantly change evoked cortical firing rates (95% HDI overlaps 0). Basal firing rates of neurons were significantly above zero (95% HDI does not overlap 0, MAP estimate = 2.2). C. Evoked single unit spike train information increases as INS energy increases.

Figure 3.

A. Evoked cortical firing activity was classified into onset, onset-sustained, sustained, and offset classes. Any response which showed an offset inhibition resulting in basal firing rate <5% prestimulus firing rate was given an inhibition designation (top left, Onset for example). B. Decomposition of response classes into the top 3 principal components show that these classes exist across a continuum.

Figure 4:

Joint peristimulus time histogram analysis reveals INS thalamocortical recruitment is spatially constrained. A. Schematic of JPSTH analysis. Covariance maps were first calculated between the two PSTHs under test. Covariance maps represent the joint activity of two neurons due to the INS stimulus directly. Subtracting the covariance map from the joint histogram generates the JPSTH, a measurement of correlated activity of the neural network in response to the stimulus. Creating a histogram of the main diagonal of the JPSTH creates a coincidence histogram of total synchrony of the two neurons. Finally, cross correlograms create a statistic of connectivity of the two neurons. Covariance and JPSTH joint histograms were smoothed by a 2D gaussian filter for visualization purposes, but full calculations were performed on raw joint histograms. B. INS-induced correlations show that lateral spread of activation in cortex from thalamic INS were constrained to ≤ 1500 μm, with 90% of responses constrained to ≤ 1000 μm. Laser energies < 1mJ limited lateral spread to ≤ 1250 μm. C. Pairwise JPTHs, measuring post-stimulation induced connectivity show lateral spreads limited to ≤ 1250 μm across all applied energies and ≤ 1000 μm for stimulus energies < 1 mJ. All correlations and JPSTHs shown were statistically significant (p < 0.05) after permutation testing.

Figure 5:

SpikerNet, a deep reinforcement learning based closed loop control system. A. Schematic of SpikerNet operation, which utilizes TD3 reinforcement learning. The state is representative of a response as recorded from the electrode environment. The agent is the set of all safe stimulation parameters. B. SpikerNet is able to find arbitrary neural firing patterns through repeated iterations of stimulation through the environment. C. SpikerNet partakes in search and targeting behavior to find target responses and to learn stimulation parameters which best drive the neural environment to target state. D. Example evoked responses during SpikerNet search and learning show a wide variety of firing classes are evoked during algorithm search. While fits were calculated around the window of evoked activity, more complex multi-peaked and offset responses were observed (Trial 12, 22, 26).