Limits...
A brain-machine interface for control of medically-induced coma.

Shanechi MM, Chemali JJ, Liberman M, Solt K, Brown EN - PLoS Comput. Biol. (2013)

Bottom Line: Medically-induced coma is a drug-induced state of profound brain inactivation and unconsciousness used to treat refractory intracranial hypertension and to manage treatment-resistant epilepsy.The median performance error for the BMI was 3.6%, the median bias was -1.4% and the overall posterior probability of reliable control was 1 (95% Bayesian credibility interval of [0.87, 1.0]).A BMI can maintain reliable and accurate real-time control of medically-induced coma in a rodent model suggesting this strategy could be applied in patient care.

View Article: PubMed Central - PubMed

Affiliation: School of Electrical and Computer Engineering, Cornell University, Ithaca, New York, United States of America ; Department of Electrical Engineering and Computer Science, University of California, Berkeley, California, United States of America.

ABSTRACT
Medically-induced coma is a drug-induced state of profound brain inactivation and unconsciousness used to treat refractory intracranial hypertension and to manage treatment-resistant epilepsy. The state of coma is achieved by continually monitoring the patient's brain activity with an electroencephalogram (EEG) and manually titrating the anesthetic infusion rate to maintain a specified level of burst suppression, an EEG marker of profound brain inactivation in which bursts of electrical activity alternate with periods of quiescence or suppression. The medical coma is often required for several days. A more rational approach would be to implement a brain-machine interface (BMI) that monitors the EEG and adjusts the anesthetic infusion rate in real time to maintain the specified target level of burst suppression. We used a stochastic control framework to develop a BMI to control medically-induced coma in a rodent model. The BMI controlled an EEG-guided closed-loop infusion of the anesthetic propofol to maintain precisely specified dynamic target levels of burst suppression. We used as the control signal the burst suppression probability (BSP), the brain's instantaneous probability of being in the suppressed state. We characterized the EEG response to propofol using a two-dimensional linear compartment model and estimated the model parameters specific to each animal prior to initiating control. We derived a recursive Bayesian binary filter algorithm to compute the BSP from the EEG and controllers using a linear-quadratic-regulator and a model-predictive control strategy. Both controllers used the estimated BSP as feedback. The BMI accurately controlled burst suppression in individual rodents across dynamic target trajectories, and enabled prompt transitions between target levels while avoiding both undershoot and overshoot. The median performance error for the BMI was 3.6%, the median bias was -1.4% and the overall posterior probability of reliable control was 1 (95% Bayesian credibility interval of [0.87, 1.0]). A BMI can maintain reliable and accurate real-time control of medically-induced coma in a rodent model suggesting this strategy could be applied in patient care.

Show MeSH

Related in: MedlinePlus

Comparison of the bounded LQR and MPC strategies with upper-bound constraints on the drug infusion rates.In each subfigure, the top panel shows the closed-loop controlled BSP traces using the bounded LQR control strategy and using the MPC strategy with various time horizons,  time samples (seconds). The bottom panel shows the corresponding drug infusion rates. In addition to being non-negative, here the drug infusion rate is required to be less than 2.4 mg/min. Here we have shown two example permutations of the target levels but the bounded LQR and the MPC drug infusion rates converge with increasing  in all cases.
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC3814408&req=5

pcbi-1003284-g005: Comparison of the bounded LQR and MPC strategies with upper-bound constraints on the drug infusion rates.In each subfigure, the top panel shows the closed-loop controlled BSP traces using the bounded LQR control strategy and using the MPC strategy with various time horizons, time samples (seconds). The bottom panel shows the corresponding drug infusion rates. In addition to being non-negative, here the drug infusion rate is required to be less than 2.4 mg/min. Here we have shown two example permutations of the target levels but the bounded LQR and the MPC drug infusion rates converge with increasing in all cases.

Mentions: We also tested the model predictive controller with various time horizons, . In the model predictive controller, we impose the constraints on the control inputs (i.e., drug infusion rates) explicitly in the formulation and thus find the constrained (approximately) optimal solution. Since our goal is to compare the bounded LQR and MPC control strategies in this set of simulations, we assume that both controllers know the BSP perfectly at each time (i.e., we use the true as feedback in the controller). We compare the MPC drug infusion rate with the bounded LQR infusion rate in Figure 4, where the constraint is positivity on the drug infusion rate. As we increase the optimization horizon, the two solutions converge. This shows that, in this problem, solving the unconstrained LQR and then bounding it is approximately optimal. The controlled BSP in Figure 3 is noisier than in Figure 4 because in the former the BSP is estimated from a stochastic binary time-series emulating the segmented EEG (Figure 1b) and in the latter BSP is assumed to be perfectly known to the controllers. We can also show that, similarly, when an upper-bound on the drug infusion rate is desired, the two solutions again converge (Figure 5). It is important to note, however, that in our problem no constraints are placed on the state. Our recursive Bayesian estimator combined with the implemented real-time MPC can extend our framework to solving more complex problems with constraints also on the state variables, such as blood pressure (see Discussion).


A brain-machine interface for control of medically-induced coma.

Shanechi MM, Chemali JJ, Liberman M, Solt K, Brown EN - PLoS Comput. Biol. (2013)

Comparison of the bounded LQR and MPC strategies with upper-bound constraints on the drug infusion rates.In each subfigure, the top panel shows the closed-loop controlled BSP traces using the bounded LQR control strategy and using the MPC strategy with various time horizons,  time samples (seconds). The bottom panel shows the corresponding drug infusion rates. In addition to being non-negative, here the drug infusion rate is required to be less than 2.4 mg/min. Here we have shown two example permutations of the target levels but the bounded LQR and the MPC drug infusion rates converge with increasing  in all cases.
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC3814408&req=5

pcbi-1003284-g005: Comparison of the bounded LQR and MPC strategies with upper-bound constraints on the drug infusion rates.In each subfigure, the top panel shows the closed-loop controlled BSP traces using the bounded LQR control strategy and using the MPC strategy with various time horizons, time samples (seconds). The bottom panel shows the corresponding drug infusion rates. In addition to being non-negative, here the drug infusion rate is required to be less than 2.4 mg/min. Here we have shown two example permutations of the target levels but the bounded LQR and the MPC drug infusion rates converge with increasing in all cases.
Mentions: We also tested the model predictive controller with various time horizons, . In the model predictive controller, we impose the constraints on the control inputs (i.e., drug infusion rates) explicitly in the formulation and thus find the constrained (approximately) optimal solution. Since our goal is to compare the bounded LQR and MPC control strategies in this set of simulations, we assume that both controllers know the BSP perfectly at each time (i.e., we use the true as feedback in the controller). We compare the MPC drug infusion rate with the bounded LQR infusion rate in Figure 4, where the constraint is positivity on the drug infusion rate. As we increase the optimization horizon, the two solutions converge. This shows that, in this problem, solving the unconstrained LQR and then bounding it is approximately optimal. The controlled BSP in Figure 3 is noisier than in Figure 4 because in the former the BSP is estimated from a stochastic binary time-series emulating the segmented EEG (Figure 1b) and in the latter BSP is assumed to be perfectly known to the controllers. We can also show that, similarly, when an upper-bound on the drug infusion rate is desired, the two solutions again converge (Figure 5). It is important to note, however, that in our problem no constraints are placed on the state. Our recursive Bayesian estimator combined with the implemented real-time MPC can extend our framework to solving more complex problems with constraints also on the state variables, such as blood pressure (see Discussion).

Bottom Line: Medically-induced coma is a drug-induced state of profound brain inactivation and unconsciousness used to treat refractory intracranial hypertension and to manage treatment-resistant epilepsy.The median performance error for the BMI was 3.6%, the median bias was -1.4% and the overall posterior probability of reliable control was 1 (95% Bayesian credibility interval of [0.87, 1.0]).A BMI can maintain reliable and accurate real-time control of medically-induced coma in a rodent model suggesting this strategy could be applied in patient care.

View Article: PubMed Central - PubMed

Affiliation: School of Electrical and Computer Engineering, Cornell University, Ithaca, New York, United States of America ; Department of Electrical Engineering and Computer Science, University of California, Berkeley, California, United States of America.

ABSTRACT
Medically-induced coma is a drug-induced state of profound brain inactivation and unconsciousness used to treat refractory intracranial hypertension and to manage treatment-resistant epilepsy. The state of coma is achieved by continually monitoring the patient's brain activity with an electroencephalogram (EEG) and manually titrating the anesthetic infusion rate to maintain a specified level of burst suppression, an EEG marker of profound brain inactivation in which bursts of electrical activity alternate with periods of quiescence or suppression. The medical coma is often required for several days. A more rational approach would be to implement a brain-machine interface (BMI) that monitors the EEG and adjusts the anesthetic infusion rate in real time to maintain the specified target level of burst suppression. We used a stochastic control framework to develop a BMI to control medically-induced coma in a rodent model. The BMI controlled an EEG-guided closed-loop infusion of the anesthetic propofol to maintain precisely specified dynamic target levels of burst suppression. We used as the control signal the burst suppression probability (BSP), the brain's instantaneous probability of being in the suppressed state. We characterized the EEG response to propofol using a two-dimensional linear compartment model and estimated the model parameters specific to each animal prior to initiating control. We derived a recursive Bayesian binary filter algorithm to compute the BSP from the EEG and controllers using a linear-quadratic-regulator and a model-predictive control strategy. Both controllers used the estimated BSP as feedback. The BMI accurately controlled burst suppression in individual rodents across dynamic target trajectories, and enabled prompt transitions between target levels while avoiding both undershoot and overshoot. The median performance error for the BMI was 3.6%, the median bias was -1.4% and the overall posterior probability of reliable control was 1 (95% Bayesian credibility interval of [0.87, 1.0]). A BMI can maintain reliable and accurate real-time control of medically-induced coma in a rodent model suggesting this strategy could be applied in patient care.

Show MeSH
Related in: MedlinePlus