Implantable UWB Antennas for Wireless Data Comunnications

    Ultra-wideband (UWB) short-range communication systems are valuable in medical technology, particularly for implanted devices, due to their low-power consumption, low cost, small size and high data rates. Monitoring of neural responses in the brain requires high data rate (800 kb/s per neural sensor), and we target a system supporting a large number of sensors, in particular, aggregate transmission above 430 Mb/s (~512 sensors). Knowledge of channel behavior is required to determine the maximum allowable power to 1) respect ANSI guidelines for avoiding tissue damage and 2) respect FCC guidelines on unlicensed transmissions. We exploited a realistic model of the biological channel to inform the design of antennas for the implanted transmitter and the external receiver under these requirements.

    Modeling of biological channel for wireless UWB communications in implantable brain computer interfaces

    Wireless Power Transmission

    The need to study freely moving animals for uninterrupted periods of times has created a growing demand for high-efficiency wireless power transmission systems. We present a novel resonance-based multicoil array structure to wirelessly power up smart neurological research systems. The proposed array consists of a novel multicoil inductive link, which primary resonator is made of several identical coil elements connected in parallel, and tiled in an array. The proposed approach 1) can deliver power with superior efficiency over longer separation distances, 2) can naturally track the receiver position and localize transmitted power through nearby coil array elements without the need for complex control and detection circuitry, and 3) can either accommodate short or long range power transmission applications. An example video is available here.

    A smart experimental chamber with uniform wireless power transmission in 3D for enabling long term nerological experiments with freely moving models.

    Wireless Multi-Channel Optogenetics Headstage

    This project aims to deliver a miniature optogenetics headstage for wirelessly stimulating the brain of rodents with four implanted LEDs and at the same time recording electrophysiological data from a 32-channel readout. The headstage is powered using a small battery, and is built using inexpensive commercial off-the-shelf electronic components, including a low-power RF transceiver and an FPGA. This headstage provides the researcher with flexible optical stimulation patterns. The neural data (received from 32 channels) is processed and compressed in real-time. This allows efficient use of radio channel and also the processed data (spikes) can be used to closed-loop brain stimulation. 

    Block diagram of a head mountable system with optical stimulation and multichannel neural recording capability and 3D Model of the headstage prototype

    DNA Assembly

    Since a few years now, DNA sequencing and assembling gets lots of attention in health science. Although sequencing has got somewhat accessible with the more recent technologies, the assembling still requires a terrific amount of computation power for larger genomes. Supercomputers are thus used to complete the task, but unfortunately, they are not accessible to everyone, including doctors. The projet aims to rewrite a currently used assembly algorithm to make use of many different hardware platforms with OpenCL. This algorithm uses DeBruijn graphs to store what are called « kmers »; small portions of ADN. The solution to the problem is the traversal of the graph using a set of heuristics to determine the most confident outcome.

    Multi-Channel Analog Front-End

    Neural recording systems featuring several parallel readout channels enable neuroscientists to study brain Neuro-dynamics in-vivo. Such invaluable information is critical to study and understand biological neural networks. A high-resolution multichannel analog front-end (AFE) is key to interface with modern implantable microelectrode arrays, which presents hundreds of recording sites. Indeed, a neural recording AFE system must present a tremendous number of Biopotential measurement channels to capture the activity of large groups of neurons. In contrast, such an AFE needs to be small and low-power to be attached or implanted inside the body without harming biological tissues. Additionally, reading circuits must present very low-noise to address very low-amplitude neural signals (i.e. 50 μV – 500 μV), and provide some mechanism to block microelectrode potentials, which can saturate low-noise amplifiers.