Page Comparison

Daniel BD Rubin (Deactivated) , Yulia Sandamirskaya (Deactivated)

(Invited Guest speaker)

Christoph Guger, g. tech medical engineering GmbH, Schleidberg, Austria

Current and future applications of brain-computer interfaces

Brain-computer interfaces are extracting information in real-time from the human brain. The work with non-invasive or invasive sensors and can be used for many applications: brain assessment for patients with disorders of consciousness, rehabilitation of upper and lower limbs of stroke patients, communication interface for locked-in patients, treatment of Parkinson with deep brain stimulation and many more.

Margherita Ronchini (presenter), Hai Au Huynh, Milad Zamani, Hooman Farkhani, and Farshad Moradi, Electrical and Computer Engineering Dept., Aarhus University, Denmark

Towards an implantable Neuromorphic Computing System for online seizure detection
Neuromorphic computing systems (NCSs) could be employed as the core of brain-computer interfaces (BCIs) to process the biological signals in a neuro-inspired fashion and deliver a stimulation pattern aimed at rectifying any anomalous behavior. The underlying premise, and motive for this investigation, is that the neuromorphic device should be able to distinguish healthy from pathological activity. In particular, here, epilepsy is the model, and seizure detection is the objective. This presentation documents the journey towards the design of an embedded implantable NCS for online seizure detection from Local Field Potential signals. Initially, software simulations in NEST served an exploratory purpose, to determine the best network architecture and encoding algorithm. As an intermediate stage, the resulting network was implemented both in FPGA and DSP. Ultimately, the work led to the design of an integrated device in CMOS technology, comprising a novel encoder circuit.

Mike Ambinder, Valve Software