Computational evolution produces an improved pattern of brain stimulation to treat Parkinson’s disease in animals and patients.

Subthalamic nucleus (STN) deep brain stimulation (DBS) is an established treatment for the motor symptoms of Parkinson's disease (PD). However, the mechanisms of action of DBS are unknown. Random temporal patterns of DBS are less effective than regular DBS, but the neuronal basis for this dependence on temporal pattern of stimulation is unclear. Using a rat model of PD, we quantified the changes in behavior and single-unit activity in globus pallidus externa and substantia nigra pars reticulata during high-frequency STN DBS with different degrees of irregularity. Although all stimulus trains had the same average rate, 130-Hz regular DBS more effectively reversed motor symptoms, including circling and akinesia, than 130-Hz irregular DBS. A mixture of excitatory and inhibitory neuronal responses was present during all stimulation patterns, and mean firing rate did not change during DBS. Low-frequency (7–10 Hz) oscillations of single-unit firing times present in hemiparkinsonian rats were suppressed by regular DBS, and neuronal firing patterns were entrained to 130 Hz. Irregular patterns of DBS less effectively suppressed 7- to 10-Hz oscillations and did not regularize firing patterns. Random DBS resulted in a larger proportion of neuron pairs with increased coherence at 7–10 Hz compared with regular 130-Hz DBS, which suggested that long pauses (interpulse interval >50 ms) during random DBS facilitated abnormal low-frequency oscillations in the basal ganglia. These results suggest that the efficacy of high-frequency DBS stems from its ability to regularize patterns of neuronal firing and thereby suppress abnormal oscillatory neural activity within the basal ganglia.

Fully-implantable wireless intra-cortical Brain Machine Interfaces (iBMI) is one of the most promising next frontiers in the nascent field of neurotechnology. However, scaling the number of channels in such systems by another 10X is difficult due to power and bandwidth requirements of the wireless transmitter. One promising solution for that is to include more processing, up to the decoder, in the implant so that transmission data-rate is reduced drastically. Earlier work on neuromorphic decoder-chips only showed classification of discrete states. We present results for continuous-state decoding using a low-power neuromorphic decoder chip termed Spike-input Extreme Learning Machine (SELMA) that implements a non-linear decoder without memory and its memory based version with time delayed bins, SELMA-bins. We have compared SELMA, SELMA-bins against state of the art Steady State Kalman Filter (SSKF), a linear decoder with memory, across two different datasets involving a total of 4 non-human primates (NHPs). Results show at least a 10% (20%) or more increase in the fraction of variance accounted for (FVAF) by SELMA (SELMA-bins) over SSKF across the datasets. Estimated energy consumption comparison shows SELMA (SELMA-bins) consuming $\approx$ 9 nJ/update (23 nJ/update) against SSKF's $\approx$ 7.4 nJ/update for an iBMI with a 10 degree of freedom control. Thus, SELMA yields better performance against SSKF while consuming energy in the same range as SSKF whereas SELMA-bins performs the best with moderately increased energy consumption, albeit far less than energy required for raw-data transmission. This paves the way for reducing transmission data rates in future scaled iBMI systems.

An emerging idea in electroencephalogram motor-imagery (EEG-MI) classification is the ‘EEG-as-image’ approach. It aims to capture local EEG signal dynamics by preserving the spatial relationships of EEG channels. We hypothesize that due to the global nature of EEG modulations, a better approach is to apply global unmixing filters.