Adaptive decoding for brain-machine interfaces through Bayesian parameter updates.
Adaptive decoding for brain-machine interfaces through Bayesian parameter updates.
复制标题
DOI：
10.1162/neco_a_00207
复制
发表时间：
2011-12
影响因子：
2.9
通讯作者：
Nicolelis MA
中科院分区：
文献类型：
--
作者：
Li Z;O'Doherty JE;Lebedev MA;Nicolelis MA
脑机接口（BMIs）将大脑运动区域中记录的神经元活动转换为外部执行器的运动。神经元群体的运动表示随着时间的推移而变化，在自愿肢体运动和通过BMI控制的运动期间，由于运动学习，神经元可塑性和记录的不稳定性。为了确保长时间内准确的BMI性能，BMI解码器必须适应这些变化。我们提出了贝叶斯回归自训练方法更新的无迹卡尔曼滤波解码器的参数。这种新颖的范例使用解码器的输出来周期性地更新解码器的贝叶斯线性回归中的神经元调谐模型。我们使用两个以前已知的贝叶斯线性回归的统计公式：（1）联合公式，允许快速和精确的推理，和（2）因子分解公式，允许从更新中添加和临时省略神经元，但需要近似变分推理。为了评估这些方法，我们进行了离线重建和闭环实验与恒河猴植入皮质与微丝电极。离线重建使用记录在3只猴子的M1，S1，PMd，SMA和PP区域的数据，同时它们使用手持操纵杆控制光标。与没有更新的相同解码器相比，贝叶斯回归自训练更新显着提高了离线重建的准确性。我们对一只猴子进行了11次实时闭环实验，实验中猴子被植入了M1和S1区域。这些会议为期29天。猴子使用解码器控制光标，有更新和没有更新。这些更新保持了控制的准确性，并且不需要关于猴子手部运动的信息，关于期望运动的假设，或者作为训练信号的预期运动目标的知识。这些结果表明，贝叶斯回归自我训练可以在很长一段时间内保持BMI控制的准确性，使临床神经修复术更加可行。
Brain machine interfaces (BMIs) transform activity of neurons recorded in motor areas of the brain into movements of external actuators. Representation of movements by neuronal populations varies over time, during both voluntary limb movements and movements controlled through BMIs, due to motor learning, neuronal plasticity, and instability in recordings. To assure accurate BMI performance over long time spans, BMI decoders must adapt to these changes. We propose the Bayesian regression self-training method for updating the parameters of an unscented Kalman filter decoder. This novel paradigm uses the decoder’s output to periodically update the decoder’s neuronal tuning model in a Bayesian linear regression. We use two previously-known statistical formulations of Bayesian linear regression: (1) a joint formulation which allows fast and exact inference, and (2) a factorized formulation which allows the addition and temporary omission of neurons from updates, but requires approximate variational inference. To evaluate these methods, we performed off-line reconstructions and closed-loop experiments with Rhesus monkeys implanted cortically with micro-wire electrodes. Off-line reconstructions used data recorded in areas M1, S1, PMd, SMA, and PP of 3 monkeys while they controlled a cursor using a hand-held joystick. The Bayesian regression self-training updates significantly improved the accuracy of offline reconstructions compared to the same decoder without updates. We performed 11 sessions of real-time, closed-loop experiments with a monkey implanted in areas M1 and S1. These sessions spanned 29 days. The monkey controlled the cursor using the decoder with and without updates. The updates maintained control accuracy and did not require information about monkey hand movements, assumptions about desired movements, or knowledge of the intended movement goals as training signals. These results indicate that Bayesian regression self-training can maintain BMI control accuracy over long time periods, making clinical neuroprosthetics more viable.