Smoothness as a failure mode of bayesian mixture models in brain-machine interfaces

Siamak Yousefi, Alex Wein, Kevin C. Kowalski, Andrew G. Richardson, Lakshminarayan Srinivasan

Research output: Contribution to journalArticle

Abstract

Various recursive Bayesian filters based on reach state equations (RSE) have been proposed to convert neural signals into reaching movements in brain-machine interfaces. When the target is known, RSE produce exquisitely smooth trajectories relative to the random walk prior in the basic Kalman filter. More realistically, the target is unknown, and gaze analysis or other side information is expected to provide a discrete set of potential targets. In anticipation of this scenario, various groups have implemented RSE-based mixture (hybrid) models, which define a discrete random variable to represent target identity. While principled, this approach sacrifices the smoothness of RSE with known targets. This paper combines empirical spiking data from primary motor cortex and mathematical analysis to explain this loss in performance. We focus on angular velocity as a meaningful and convenient measure of smoothness. Our results demonstrate that angular velocity in the trajectory is approximately proportional to change in target probability. The constant of proportionality equals the difference in heading between parallel filters from the two most probable targets, suggesting a smoothness benefit to more narrowly spaced targets. Simulation confirms that measures to smooth the data likelihood also improve the smoothness of hybrid trajectories, including increased ensemble size and uniformity in preferred directions. We speculate that closed-loop training or neuronal subset selection could be used to shape the user's tuning curves towards this end.

Original languageEnglish (US)
Article number6832621
Pages (from-to)128-137
Number of pages10
JournalIEEE Transactions on Neural Systems and Rehabilitation Engineering
Volume23
Issue number1
DOIs
StatePublished - Jan 1 2015
Externally publishedYes

Fingerprint

Brain-Computer Interfaces
Failure modes
Brain
Trajectories
Angular velocity
Motor Cortex
Equations of state
Random variables
Kalman filters
Tuning
Direction compound

All Science Journal Classification (ASJC) codes

  • Neuroscience(all)
  • Biomedical Engineering
  • Computer Science Applications

Cite this

Smoothness as a failure mode of bayesian mixture models in brain-machine interfaces. / Yousefi, Siamak; Wein, Alex; Kowalski, Kevin C.; Richardson, Andrew G.; Srinivasan, Lakshminarayan.

In: IEEE Transactions on Neural Systems and Rehabilitation Engineering, Vol. 23, No. 1, 6832621, 01.01.2015, p. 128-137.

Research output: Contribution to journalArticle

Yousefi, Siamak ; Wein, Alex ; Kowalski, Kevin C. ; Richardson, Andrew G. ; Srinivasan, Lakshminarayan. / Smoothness as a failure mode of bayesian mixture models in brain-machine interfaces. In: IEEE Transactions on Neural Systems and Rehabilitation Engineering. 2015 ; Vol. 23, No. 1. pp. 128-137.
@article{809322f35e6c4bbf8ebfdd64e9ca3706,
title = "Smoothness as a failure mode of bayesian mixture models in brain-machine interfaces",
abstract = "Various recursive Bayesian filters based on reach state equations (RSE) have been proposed to convert neural signals into reaching movements in brain-machine interfaces. When the target is known, RSE produce exquisitely smooth trajectories relative to the random walk prior in the basic Kalman filter. More realistically, the target is unknown, and gaze analysis or other side information is expected to provide a discrete set of potential targets. In anticipation of this scenario, various groups have implemented RSE-based mixture (hybrid) models, which define a discrete random variable to represent target identity. While principled, this approach sacrifices the smoothness of RSE with known targets. This paper combines empirical spiking data from primary motor cortex and mathematical analysis to explain this loss in performance. We focus on angular velocity as a meaningful and convenient measure of smoothness. Our results demonstrate that angular velocity in the trajectory is approximately proportional to change in target probability. The constant of proportionality equals the difference in heading between parallel filters from the two most probable targets, suggesting a smoothness benefit to more narrowly spaced targets. Simulation confirms that measures to smooth the data likelihood also improve the smoothness of hybrid trajectories, including increased ensemble size and uniformity in preferred directions. We speculate that closed-loop training or neuronal subset selection could be used to shape the user's tuning curves towards this end.",
author = "Siamak Yousefi and Alex Wein and Kowalski, {Kevin C.} and Richardson, {Andrew G.} and Lakshminarayan Srinivasan",
year = "2015",
month = "1",
day = "1",
doi = "10.1109/TNSRE.2014.2329698",
language = "English (US)",
volume = "23",
pages = "128--137",
journal = "IEEE Transactions on Neural Systems and Rehabilitation Engineering",
issn = "1534-4320",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
number = "1",

}

TY - JOUR

T1 - Smoothness as a failure mode of bayesian mixture models in brain-machine interfaces

AU - Yousefi, Siamak

AU - Wein, Alex

AU - Kowalski, Kevin C.

AU - Richardson, Andrew G.

AU - Srinivasan, Lakshminarayan

PY - 2015/1/1

Y1 - 2015/1/1

N2 - Various recursive Bayesian filters based on reach state equations (RSE) have been proposed to convert neural signals into reaching movements in brain-machine interfaces. When the target is known, RSE produce exquisitely smooth trajectories relative to the random walk prior in the basic Kalman filter. More realistically, the target is unknown, and gaze analysis or other side information is expected to provide a discrete set of potential targets. In anticipation of this scenario, various groups have implemented RSE-based mixture (hybrid) models, which define a discrete random variable to represent target identity. While principled, this approach sacrifices the smoothness of RSE with known targets. This paper combines empirical spiking data from primary motor cortex and mathematical analysis to explain this loss in performance. We focus on angular velocity as a meaningful and convenient measure of smoothness. Our results demonstrate that angular velocity in the trajectory is approximately proportional to change in target probability. The constant of proportionality equals the difference in heading between parallel filters from the two most probable targets, suggesting a smoothness benefit to more narrowly spaced targets. Simulation confirms that measures to smooth the data likelihood also improve the smoothness of hybrid trajectories, including increased ensemble size and uniformity in preferred directions. We speculate that closed-loop training or neuronal subset selection could be used to shape the user's tuning curves towards this end.

AB - Various recursive Bayesian filters based on reach state equations (RSE) have been proposed to convert neural signals into reaching movements in brain-machine interfaces. When the target is known, RSE produce exquisitely smooth trajectories relative to the random walk prior in the basic Kalman filter. More realistically, the target is unknown, and gaze analysis or other side information is expected to provide a discrete set of potential targets. In anticipation of this scenario, various groups have implemented RSE-based mixture (hybrid) models, which define a discrete random variable to represent target identity. While principled, this approach sacrifices the smoothness of RSE with known targets. This paper combines empirical spiking data from primary motor cortex and mathematical analysis to explain this loss in performance. We focus on angular velocity as a meaningful and convenient measure of smoothness. Our results demonstrate that angular velocity in the trajectory is approximately proportional to change in target probability. The constant of proportionality equals the difference in heading between parallel filters from the two most probable targets, suggesting a smoothness benefit to more narrowly spaced targets. Simulation confirms that measures to smooth the data likelihood also improve the smoothness of hybrid trajectories, including increased ensemble size and uniformity in preferred directions. We speculate that closed-loop training or neuronal subset selection could be used to shape the user's tuning curves towards this end.

UR - http://www.scopus.com/inward/record.url?scp=84920918955&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84920918955&partnerID=8YFLogxK

U2 - 10.1109/TNSRE.2014.2329698

DO - 10.1109/TNSRE.2014.2329698

M3 - Article

VL - 23

SP - 128

EP - 137

JO - IEEE Transactions on Neural Systems and Rehabilitation Engineering

JF - IEEE Transactions on Neural Systems and Rehabilitation Engineering

SN - 1534-4320

IS - 1

M1 - 6832621

ER -