Research: Biological and Bio-inspired Supercomputers

Analog Supercomputers for Synthetic Biology and Medicine


cochleaThis project exploits the astoundingly detailed similarity between the stochastic Boltzmann equations of chemistry and the stochastic Boltzmann equations of subthreshold analog electronics [1] to attempt to create a digitally programmable analog VLSI (Very Large Scale Integration) supercomputer chipset. The supercomputing chipset has the potential for lightning-fast exact simulations of biological intracellular (gene and protein) and extracellular (hormonal, immune, neuronal, and organ) networks including their noisy or stochastic properties. Thus, it is widely useful as a design and simulation tool in synthetic biology as well as a computational tool in medicine and systems biology. Such computational tools can shed insight into the analysis and synthetic design of treatments for cancer, diabetes, auto-immune diseases, antibiotic resistance, or neuronal diseases.

  1. CYTOMORPHIC ELECTRONICS: Cell-inspired electronics for systems and synthetic biology. Chapter 24, pp. 753-786, in R. Sarpeshkar Ultra Low Power Bioelectronics: Fundamentals, Biomedical Applications, and Bio-inspired Systems, Cambridge University Press, Cambridge, February 2010.
  2. LOG DOMAIN CIRCUIT MODELS OF CHEMICAL REACTIONS: S. Mandal and R. Sarpeshkar, “Log-Domain Circuit Models of Chemical Reactions,” Proceedings of the IEEE International Symposium on Circuits and Systems (ISCAS), Taipei, Taiwan, May 2009, pp. 2697-2700.
  3. CIRCUIT MODELS OF STOCHASTIC GENETIC NETWORS: S. Mandal and R. Sarpeshkar, “Circuit Models of Stochastic Genetic Networks,” 2009 IEEE Symposium on Biological Circuits and Systems (BioCAS), Beijing, China, pp. 109-112, November 2009.
  4. ANALOG VERSUS DIGITAL: R. Sarpeshkar, “Analog Versus Digital: Extrapolating from Electronics to Neurobiology,” Neural Computation, Vol. 10, pp. 1601-1638, 1998.

Analog Ear-Inspired Supercomputer for Spectrum Analysis: The RF Cochlea

The biological inner ear or cochlea is an amazing custom analog computer capable of the equivalent of 1GFLOPS of spectral-analysis and gain-control computations with 14uW of power on a 150mV battery and a minimum detectable signal of 0.05 angstroms. It achieves such efficiency because of the clever use of an active nonlinear transmission line implemented with fluids, membranes, active piezoelectret cells, micromechanics, and electrochemistry.The cochlea has an amazingly large input dynamic range of 120dB, analyzes frequencies over a 100-fold range in carrier frequency (100Hz-10kHz), and amplifies signals at 100kHz even though its cells have time constants of 1ms. We use inspiration from the cochlea to construct an RF cochlea a fast, ultra-broadband, low-power spectrum analyzer. Instead of working with sound waves from 100Hz to 10kHz as in the audio cochlea, we work with radio waves from 100MHz to 10GHz but the principles of wave processing are similar and inspired by the biological cochlea. The actions of fluid mass in the ear are mimicked with inductors, the actions of membranes in the ear with capacitors, and the actions of outer hair cells in the ear with active RF amplifiers. Electrically, the cochlea can be modeled as an active, nonlinear, adaptive transmission line with characteristic frequencies that scale exponentially with position. Nonlinear behavior is important in the biological cochlea, particularly for signal detection in noise and gain control. We are researching how the RF cochlea may be used as a front end for universal radios, software radios, and cognitive radios and improve the detection of radio signals in noise.

Selected Publications

  1. RF COCHLEA SYSTEM: S. Mandal, S. Zhak, and R. Sarpeshkar, “A Bio-Inspired Active Radio-Frequency Silicon Cochlea,” IEEE Journal of Solid-State Circuits, Vol. 44, No. 6, pp. 1814-1828, June 2009.
  2. RF FOVEA: S. Mandal, R. Sarpeshkar, “A Bio-Inspired Cochlear Heterodyning Architecture for an RF Fovea,” IEEE Transactions on Circuits and Systems I: Regular Papers, Vol. 58, No. 7, pp. 1647 – 1660, 2011.
  3. RF COCHLEA CIRCUITS: S. Mandal, S. Zhak, and R. Sarpeshkar, “Circuits for an RF Cochlea“, Proceedings of the International Symposium on Circuits and Systems (ISCAS 2006), Kos, Greece, pp 3610–3613, May 21–24, 2006.
  4. FIRST RF COCHLEA PAPER: S. M. Zhak, S. Mandal, and R. Sarpeshkar, “A Proposal for an RF ochlea“, invited paper, Proceedings of the Asia Pacific Microwave Conference, New Delhi, India, 4 pages, December 15-18, 2004.

Analog and Bio-inspired Projects in Sensing and Computing

Analog and bio-inspired projects in the lab have led to interesting innovations and applications: A bio-inspired analog-to-digital converter built with two spiking neurons is currently at or near the state-of-the-art in energy efficiency (0.12 pJ per quantization level) for A-to-D converters. It was inspired by how spiking neurons perform pattern recognition with time rather than with voltage or current, and was the first time-based converter whose conversion time scaled linearly with bit precision instead of exponentially. A bio-inspired companding noise-reduction algorithm, inspired by the operation of the silicon and biological cochlea, led to an architecture for doing spectral analysis that has shown improvements in subjects for hearing in noise and for speech recognition in noise. A bio-inspired asynchronous interleaved sampling algorithm (AIS), inspired by the operation of winner-take-all spiking neurons, led to an algorithm for efficient low-power neural stimulation that preserves phase information well and is thus useful for encoding music or tonal languages (e.g. Chinese) in cochlear implants. A cochlear-implant chip processor that exploits this algorithm has just been built. Research in the lab on low-power wide-dynamic-range spike-based imagers and on a silicon cochlea promise improvements in imagers and audio processing in the near future. Some of Professor Sarpeshkar’s early work on motion processing in Analog VLSI was inspired by motion processing in flies.

Selected Publications

  1. BIO-INSPIRED SILICON VOCAL TRACT: K. H. Wee and L. Turicchia and R. Sarpeshkar, “An Analog Integrated-Circuit Vocal Tract,” IEEE Transactions on Biomedical Circuits and Systems, Vol.2, No. 4, pp. 316-327, 2008.
  2. BIO-INSPIRED A-TO-D: Yang, H. and R. Sarpeshkar, “A Bio-inspired Ultra-Energy-Efficient Analog-to-Digital Converter for Biomedical Applications“, IEEE Transactions on Circuits and Systems I, special issue on Life Sciences and System Applications, Vol. 53, No. 11, pp. 2349-2356, November 2006.
  3. BIO-INSPIRED ASYNCHRONOUS SAMPLING (AIS ALGORITHM): J. Sit, A. M. Simonson, A. J. Oxenham, M. A. Faltys, and R. Sarpeshkar, “A low-power asynchronous interleaved sampling algorithm for cochlear implants that encodes envelope and phase information“, IEEE Transactions on Biomedical Engineering, Vol. 54, pp. 138-149, 2007.
  4. BIO-INSPIRED ASYNCHRONOUS SAMPLING (AIS PROCESSOR): J. Sit and R. Sarpeshkar, “A Cochlear-Implant Processor for Encoding Music and Lowering Stimulation Power,” IEEE Pervasive Computing, Vol. 1, No. 7, pp. 40-48, 2008.
  5. BIO-INSPIRED COMPANDING NOISE-REDUCTION ALGORITHM: L. Turicchia and R. Sarpeshkar, “A Bio-Inspired Companding Strategy for Spectral Enhancement,” IEEE Transactions on Speech and Audio Processing, Vol. 13, No. 2, pp. 243-253, March 2005
  6. BIO-INSPIRED SILICON COCHLEA: R. Sarpeshkar, R.F. Lyon, and C.A. Mead, “A Low-Power Wide-Dynamic-Range Analog VLSI Cochlea,” Analog Integrated Circuits and Signal Processing, Vol. 13, pp. 123-151, 1997.
  7. BIO-INSPIRED VISUAL MOTION PROCESSING: R. Sarpeshkar, J. Kramer, G. Indiveri, and C. Koch, “Analog VLSI Architectures for Motion Processing: From Fundamental Limits to System Applications,&rdquo Invited Paper, Proceedings of the IEEE, Vol. 84, No. 7, pp. 969-987, 1996.
  8. HYBRID ANALOG-DIGITAL SYSTEMS INSPIRED BY SPIKING NEURONAL CIRCUITS IN BIOLOGY: R. Sarpeshkar and M. O’Halloran, “Scalable Hybrid Computation with Spikes,” Neural Computation, Vol. 14, No. 9, pp. 2003-2024, September 2002.
  9. CORTEX-INSPIRED HYBRID ANALOG-DIGITAL FEEDBACK COMPUTATION: R. Hahnloser, R. Sarpeshkar, M. Mahowald, R. Douglas, and S. Seung, ” Digital Selection and Analogue Amplification Coexist in a cortex-inspired silicon circuit,” NATURE, Cover article, Vol. 405, pp. 947-951, 22 June 2000.