NGNI: Ultra-Low Power Implantable Platform for Next Generation Neural Interfaces

Research Topics

Underpinning research related to NGNI in research topics:

See Resources/Spike Sorting Platform

Completed Project (2010-2014)

Research Team: Yan Liu, Song Luan, Sivylla Paraskevopoulou, Deren Barsakcioglu, Amir Eftekhar, Timothy Constandinou
Collaborators: Andrew Jackson (Newcastle), Rodrigo Quian Quiroga (Leicester) 
Funding: Engineering and Physical Sciences Research Council (EPSRC) EP/I000569/1

Next Generation Neural Interfaces Concept
Wireless, multi-channel neural recording interface with on-node spike sorting for real-time motor prosthetic control

For over half a century scientists have recorded the tiny electrical potentials generated by neurons in the brains of awake animals performing specific behaviours, using large racks of power-hungry equipment. These experiments have yielded profound insights into how sensory information is represented and transformed by the brain into the signals that control purposeful movements, as well as revealing how this complex system is affected by neurological injuries and disease. However, until recently the therapeutic avenues available to neurologists have been limited to gross interventions such as systemic drug applications and neurosurgical lesions.

In recent years, small electronic devices have been developed that deliver specific patterns of stimulation via small electrodes implanted in the nervous system. Devices such as Deep Brain Stimulators and Cochlear Implants have helped many thousands of patients worldwide. The next generation of neural implants will use similar electrodes to detect the activity of neurons, paving the way for new treatments for conditions that currently weigh a heavy clinical burden. For example, by using the activity of neurons in motor areas of the brain to control electrical stimulation of muscles, it is possible that voluntary movements could be restored to patients paralysed by spinal cord injuries. However, despite considerable advances in electrode technologies, our ability to interface digital microelectronics with the brain at the level of individual neurons is at present severely limited. Each electrode detects the signal from multiple cells in its vicinity, and the small, brief 'spike' events they generate can be hard to distinguish beneath the background noise.

To address this problem we assembled a cross-disciplinary team with expertise in three key areas: the computational algorithms required to detect and sort spike events, low power integrated electronics to perform real-time, reliable spike identification, and the techniques to record long-term activity from the brain using neural implants in order to evaluate real-world performance. The project successfully delivered a platform technology for converting the raw signal from electrodes into a stream of identified spike events suitable for subsequent processing by conventional digital microelectronics.

Outcomes

The Next Generation Neural Interfaces project has successfully established a suite of algorithms and methods, as well as both software and hardware tools and resources for neural recording, realtime spike sorting and implantable devices.

Template-based Spike Sorting: We established a workflow for on-node, realtime spike sorting. This combines a computationally-intensive "workstation-calibrated" clustering method (WaveClus) for training with more hadware efficient template-based classification. Specific outcomes include:

  • Analogue front end parameters (i.e. noise, amplifier bandwidth, gain, filter order, type, cut-off frequencies, data converter specifications, etc) were optimised for resource efficient (low power) spike sorting performance (classification sensitivity).
  • Sensitivity of the various parameters and methods used in template matching itself (i.e. alignment method, distance measure, template window, alignment offset, sample rate, resolution, etc) were established with respect to spike sorting performance with a view to hardware implementability.

Tools and Resources: We developed a number of tools and resources (both software and hardware) for experimental neuroscience. These include:

  • A scalable neural recording platform with on-node spike sorting based on template matching.
  • Analogue front end behavioural modelling tool (Matlab based).

New algorithms and methods for spike sorting: Through this project we identified a number of new feature sets (e.g. First and Second Derivative Extrema), classification/clustering methods (e.g. Hierarchical Adaptive Means), and calibration-free methods for hardware-efficient spike sorting towards 1,000 channel on-node processing.

Publications

2018

  • Luan S, Williams I, Maslik M, Liu Y, De Carvalho F, Jackson A, Quiroga RQ, Constandinou TG, 2018, Compact standalone platform for neural recording with real-time spike sorting and data logging, JOURNAL OF NEURAL ENGINEERING, Vol: 15, ISSN: 1741-2560

2014

  • Williams, I., Luan, S., Jackson, A., & Constandinou, T. G. (2015). A Scalable 32 Channel Neural Recording and Real-time FPGA Based Spike Sorting System. In Biomedical Circuits and Systems (BioCAS), 2015 IEEE International Conference on.
  • Leene, L. B., & Constandinou, T. G. (2014). Ultra-low power design strategy for two-stage amplifier topologies. Electronics Letters, 50(8), 583.
  • Luan, S., & Constandinou, T. G. (2014). A charge-metering method for voltage-mode neural stimulation. Journal of neuroscience methods, 224, 39-47.
  • Zheng, L., Leene, L. B., Liu, Y., & Constandinou, T. G. (2014). An adaptive 16/64 kHz, 9-bit SAR ADC with peak-aligned sampling for neural spike recording. In Circuits and Systems (ISCAS), 2014 IEEE International Symposium on (pp. 2385-2388).
  • Barsakcioglu, D. Y., Liu, Y., Bhunjun, P., Navajas, J., Eftekhar, A., Jackson, A., ... & Constandinou, T. G. (2014). An Analogue Front-End Model for Developing Neural Spike Sorting Systems. Biomedical Circuits and Systems, IEEE Transactions on, 8(2), 216-227.
  • Paraskevopoulou, S. E., Wu, D., Eftekhar, A., & Constandinou, T. G. (2014). Hierarchical Adaptive Means (HAM) clustering for hardware-efficient, unsupervised and real-time spike sorting. Journal of neuroscience methods, 235, 145-156.
  • Navajas, J., Barsakcioglu, D. Y., Eftekhar, A., Jackson, A., Constandinou, T. G., & Quiroga, R. Q. (2014). Minimum requirements for accurate and efficient real-time on-chip spike sorting. Journal of neuroscience methods, 230, 51-64.
  • Luan, S., Williams, I., Nikolic, K., & Constandinou, T. G. (2014). Neuromodulation: present and emerging methods. Frontiers in neuroengineering, 7.
  • Eftekhar, A., Juffali, W., El-Imad, J., Constandinou, T. G., & Toumazou, C. (2014). Ngram-Derived Pattern Recognition for the Detection and Prediction of Epileptic Seizures.
  • Williams, I., & Constandinou, T. G. (2014). Computationally efficient modeling of proprioceptive signals in the upper limb for prostheses: a simulation study. Frontiers in neuroscience, 8.

2013

  • Barsakcioglu, D. Y., Eftekhar, A., & Constandinou, T. G. (2013). Design optimisation of front-end neural interfaces for spike sorting systems. In Circuits and Systems (ISCAS), 2013 IEEE International Symposium on (pp. 2501-2504).
  • Williams, I., & Constandinou, T. G. (2013). An energy-efficient, dynamic voltage scaling neural stimulator for a proprioceptive prosthesis. Biomedical Circuits and Systems, IEEE Transactions on, 7(2), 129-139.
  • Koutsos, E., Paraskevopoulou, S. E., & Constandinou, T. G. (2013). A 1.5 μW NEO-based spike detector with adaptive-threshold for calibration-free multichannel neural interfaces. In Circuits and Systems (ISCAS), 2013 IEEE International Symposium on (pp. 1922-1925).
  • Leene, L. B., Liu, Y., & Constandinou, T. G. (2013). A compact recording array for neural interfaces. In Biomedical Circuits and Systems Conference (BioCAS), 2013 IEEE (pp. 97-100).
  • Leene, L. B., Luan, S., & Constandinou, T. G. (2013). A 890fJ/bit UWB transmitter for SOC integration in high bit-rate transcutaneous bio-implants. In Circuits and Systems (ISCAS), 2013 IEEE International Symposium on (pp. 2271-2274).

2012

  • Guilvard, A., Eftekhar, A., Luan, S., Toumazou, C., & Constandinou, T. G. (2012). A fully-programmable neural interface for multi-polar, multi-channel stimulation strategies. In Circuits and Systems (ISCAS), 2012 IEEE International Symposium on (pp. 2235-2238).
  • Luan, S., & Constandinou, T. G. (2012). A novel charge-metering method for voltage mode neural stimulation. In Circuits and Systems (ISCAS), 2012 IEEE International Symposium on (pp. 2239-2242).
  • Mirza, K. B., Luan, S., Eftekhar, A., & Constandinou, T. G. (2012). Towards a fully-integrated solution for capacitor-based neural stimulation. In Circuits and Systems (ISCAS), 2012 IEEE International Symposium on (pp. 2243-2246).
  • Haaheim, B., & Constandinou, T. G. (2012). A sub-1µW, 16kHz current-mode SAR-ADC for single-neuron spike recording. In Circuits and Systems (ISCAS), 2012 IEEE International Symposium on (pp. 2957-2960).
  • Paraskevopoulou, S. E., & Constandinou, T. G. (2012). An ultra-low-power front-end neural interface with automatic gain for uncalibrated monitoring. In Circuits and Systems (ISCAS), 2012 IEEE International Symposium on (pp. 193-196). IEEE.

2011

  • Paraskevopoulou, S. E., & Constandinou, T. G. (2011). A sub-1µW neural spike-peak detection and spike-count rate encoding circuit. In Biomedical Circuits and Systems Conference (BioCAS), 2011 IEEE (pp. 29-32).
  • Sole, M., Sanni, A., Vilchesy, A., Toumazou, C., & Constandinou, T. G. (2011). A bio-implantable platform for inductive data and power transfer with integrated battery charging. In Circuits and Systems (ISCAS), 2011 IEEE International Symposium on (pp. 2605-2608).

2010

  • Eftekhar, A., Paraskevopoulou, S. E., & Constandinou, T. G. (2010). Towards a next generation neural interface: Optimizing power, bandwidth and data quality. In Biomedical Circuits and Systems Conference (BioCAS), 2010 IEEE (pp. 122-125).