A Feedback Information-Theoretic Approach to the Design of Brain–Computer Interfaces

Omar, Cyrus; Akce, Abdullah; Johnson, M. J.; Bretl, Timothy; Ma, Rui; Maclin, Edward L.; McCormick, Martin; Coleman, Todd P.

doi:10.1080/10447318.2011.535749

Cited by 38 publications

(33 citation statements)

References 36 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…From the neuroscientific standpoint, this approach can be viewed as directing the process of operant conditioning (reward-based learning, [26]) towards neural signal characteristics that favor high quality control. Building on initial work in this area [2], [13], [14], [16], with connections to an expansive related multi-disciplinary literature, the approach to neural shaping outlined in this paper suggests one path in the near term for experiments to demonstrate and break local Pareto-optimal solutions [19] resulting from existing BCI algorithms based on neural decoders.…”

Section: Discussionmentioning

confidence: 99%

“…BCI algorithms govern how computer states change with neural activity and what feedback is sent to the user. Although BCI can be used for typing, drawing, and other activities based on discretevalued user commands [2], we focus on user control of movement based on continuous-valued user commands.…”

Section: Introductionmentioning

confidence: 99%

“…In one demonstration of this approach, the posterior matching scheme [15] was employed in the context of discrete-valued (binary) user commands to specify a two-dimensional path of movement [2]. In this case, adopting the encoding scheme did not likely require significant user training.…”

Section: Introductionmentioning

confidence: 99%

“…Separate work antecedant to the newer adaptive decoders proposed a comprehensive framework for the BCI algorithm as the joint selection of encoder and decoder [2], [13], [14]. In one demonstration of this approach, the posterior matching scheme [15] was employed in the context of discrete-valued (binary) user commands to specify a two-dimensional path of movement [2].…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Neural shaping with joint optimization of controller and plant under restricted dynamics

Srinivasan

2014

2014 Information Theory and Applications Workshop (ITA)

View full text Add to dashboard Cite

Abstract-The prototypical brain-computer interface (BCI) algorithm translates brain activity into changes in the states of a computer program, for typing or cursor movement. Most approaches use neural decoding which learns how the user has encoded their intent in their noisy neural signals. Recent adaptive decoders for cursor movement improved BCI performance by modeling the user as a feedback controller; when this model accounts for adaptive control, the neural decoder is termed co-adaptive. This recent collection of control-inspired neural decoding strategies disregards a major antecedent conceptual realization, whereby the user could be induced to adopt an encoding strategy (control policy) such that the encoder-decoder pair (or equivalently, controller-plant pair) is optimal under a desired cost function.We call this alternate conceptual approach neural shaping, in contradistinction to neural decoding. Previous work illuminated the general form of optimal controller-plant pair under a cost representing information gain. For BCI applications requiring the user to issue discrete-valued commands, the information-gainoptimal pair, based on the posterior matching scheme, can be user-friendly.In this paper, we discuss the application of neural shaping to cursor control with continuous-valued states based on continuous-valued user commands. We examine the problem of jointly optimizing controller and plant under quadratic expected cost and restricted linear plant dynamics. This simplification reduces joint controller-plant selection to a static optimization problem, similar to approaches in structural engineering and other areas. This perspective suggests that recent BCI approaches that alternate between adaptive neural decoders and static neural decoders could be local Pareto-optimal, representing a suboptimal iterative-type solution to the optimal joint controllerplant problem.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Neural shaping with joint optimization of controller and plant under restricted dynamics

Srinivasan

2014

2014 Information Theory and Applications Workshop (ITA)

View full text Add to dashboard Cite

show abstract

“…Separate work by Gilja et al (2010), and subsequently Dangi et al (2011) and Orsborn et al (2012) examined modified "cursorGoal" state equations that incorporate the current state of the actuator (an on-screen pointer) in the context of a recursive Bayesian decoder. Newer work along these lines incorporates feedback information theory (Omar et al, 2011) and team decision theory (Kim & Coleman, 2011) in the closed-loop design of BMI algorithms. Most recently, Golub et al (2011) proposed a BMI algorithm that demonstrates particularly poor open-loop decoding error but enhanced closed-loop control relative to other competing methods that have better open-loop decoding error.…”

Section: Implications For Bmi Algorithmmentioning

confidence: 99%

Stochastic Optimal Control as a Theory of Brain-Machine Interface Operation

Lagang

Srinivasan

2013

Neural Computation

View full text Add to dashboard Cite

The closed-loop operation of brain-machine interfaces (BMI) provides a framework to study the mechanisms behind neural control through a restricted output channel, with emerging clinical applications to stroke, degenerative disease, and trauma. Despite significant empirically driven improvements in closed-loop BMI systems, a fundamental, experimentally validated theory of closed-loop BMI operation is lacking. Here we propose a compact model based on stochastic optimal control to describe the brain in skillfully operating canonical decoding algorithms. The model produces goal-directed BMI movements with sensory feedback and intrinsically noisy neural output signals. Various experimentally validated phenomena emerge naturally from this model, including performance deterioration with bin width, compensation of biased decoders, and shifts in tuning curves between arm control and BMI control. Analysis of the model provides insight into possible mechanisms underlying these behaviors, with testable predictions. Spike binning may erode performance in part from intrinsic control-dependent constraints, regardless of decoding accuracy. In compensating decoder bias, the brain may incur an energetic cost associated with action potential production. Tuning curve shifts, seen after the mastery of a BMI-based skill, may reflect the brain's implementation of a new closed-loop control policy. The direction and magnitude of tuning curve shifts may be altered by decoder structure, ensemble size, and the costs of closed-loop control. Looking forward, the model provides a framework for the design and

show abstract

Interface Design Challenge for Brain-Computer Interaction

Hill

Brunner

Vaughan

2011

Foundations of Augmented Cognition. Directing the Future of Adaptive Systems

View full text Add to dashboard Cite

Great things can be achieved even with very low bandwidth. Stephen Hawking has been able to break new ground in theoretical physics just by twitching his hand and cheek. Jean-Dominique Bauby was able to write a bestselling memoir by blinking one eyelid. By reading and decoding "brain-waves", the field of brain-computer interfacing (BCI) is poised to open up the possibility of such expression, even for people who can no longer move a single muscle. A BCI still requires an HCI front-end to be of practical use, but many currentlyused HCIs do not adequately address limitations on the typical target user's input (e.g., limited eye movement leading to poor spatial vision) or output (e.g. variable delays, and false positives/negatives, in "pressing the button"). In this symposium, BCI experts will present their view of the challenges arising from these limitations. The HCI community is invited to participate in a competition to provide the best solutions.

show abstract

A Feedback Information-Theoretic Approach to the Design of Brain–Computer Interfaces

Cited by 38 publications

References 36 publications

Neural shaping with joint optimization of controller and plant under restricted dynamics

Neural shaping with joint optimization of controller and plant under restricted dynamics

Stochastic Optimal Control as a Theory of Brain-Machine Interface Operation

Interface Design Challenge for Brain-Computer Interaction

Contact Info

Product

Resources

About