Program Day 2 - January 10, 2023

Bio: Anand directs the Integrated Systems Laboratory in the Elmore Family School of Electrical and Computer Engineering at Purdue. His group’s research spans various topics in VLSI and Computer Engieering, including System-on-chip design, domain-specific architecture, computing with nanoscale post-CMOS devices, and heterogeneous parallel computing. He is Associate Director of the SRC/DARPA Center for Brain-inspired Computing and Co-Director of the Center for a Secure Microelectronics Ecosystem. Before joining Purdue, Anand was a Senior Researcher at NEC Laboratories America and held a visiting position at the Department of Electrical Engineering, Princeton University.

Title: Narrowing the energy efficiency gap between artificial and natural intelligence

Abstract: Improvements in computing performance have been a major enabler of the advances in AI over the past decade. However, we are at a point where demands from future AI workloads will far outpace expected improvements in hardware, threatening to greatly impede continued progress in the field of AI. This talk will review the challenges posed by AI workloads across the computing spectrum, examine the often-cited efficiency gap between artificial intelligence and biological intelligence, and provide a possible roadmap to narrowing this gap, including in-memory computing, algorithm-hardware co-design and neuromorphic computing.

Session Chair: Chetan Singh Thakur

Bio: Arindam Basu received the B.Tech. and M.Tech. degrees in ECE from the I.I.T, Kharagpur, India, and the M.S. degree inMathematics and the Ph.D. degree in ECE from the Georgia Institute of Technology, Atlanta, GA, USA. He is currently a Professor with the Department of EE, City University of Hong Kong and was a tenured faculty at NTU, Singapore previously.

Abstract: Dr. Basu was included in Georgia Tech Alumni Association’s 40 under 40 list in 2021 and was awarded the MIT Technology Review’s TR35 Asia Pacific Award in 2012. He also received the Prime Minister of India Gold Medal from I.I.T Kharagpur in 2005. He and his students have received several best paper awards and nominations in IEEE conferences.
He has served as IEEE CAS Distinguished Lecturer from 2016 to 2017 and currently serves IEEE in various roles such as TC Chair, Associate editor of journals etc.

Title: Neuromorphic technologies for Brain-Machine Interfaces

Neuromorphic electronics take inspiration from the brain to develop circuits and systems with similar energy and area efficiencies. While they have become very popular for different edge AI applications, these technologies have major benefits for biomedical applications with a natural fit to brain-machine interfaces (BMI). In this talk, I will outline some of our group’s work in developing neuromorphic intention decoder integrated circuits with very low energy/area footprint suitable for implantation. I will also show decoding algorithms designed jointly with the hardware inspired by neuronal population coding which can be trained rapidly to adapt to the non-stationary nature of the acquired signals. Lastly, I will talk about some future directions such as integration with sensory feedback from neuromorphic e-skin and self-adapting decoders.

Session Chair: Chetan Singh Thakur

Bio:  Dr. Sahil Shah is an Assistant Professor in the Department of Electrical and Computer Engineering at the University of Maryland, College Park. He officially joined the UMD ECE Department in Spring 2021. His area of expertise is low-power analog and mixed-signal systems for energy-efficient computation. His lab investigates and designs low-power systems that can compute efficiently in a low-resource environment, such as implantable or wearable platforms. Prior to his arrival at the University of Maryland, Sahil was a postdoctoral associate in the department of Electrical Engineering at California Institute of Technology. At Caltech, he pursued research on developing robust brain-machine interface for enabling patients to control prosthetic devices. He received his PhD in Electrical Engineering from Georgia Institute of Technology in 2018 where he developed reconfigurable mixed-signal neural networks for monitoring vital and physiological signals. His research interests fall into three major areas: Energy-Efficient integrated circuits, Embedded Machine learning, and Bio-Sensing and Monitoring. Sahil’s long term goal is to develop robust and energy-efficient devices that will equip physicians with tools to make better diagnosis, tailor rehabilitation process for patients and technology that will help us better understand physiological and neurological activity.

Title: Energy-Efficient Computing for Robust Brain Machine Interface

Abstract: This talk focuses on designing and developing energy-efficient architectures and algorithms for robust and accurate Brain Machine interfaces (BMI). A BMI is a technological system that records neural signals and maps them to control prosthetic device parameters. Thus, enabling patients with motor impairments to interact seamlessly with their environment. The talk will focus on intracortical BMIs in human participants and animal models (where electrodes are implanted in specific cortical areas). The talk will cover the general challenges involved in developing the implanted BMIs. Further, the talk will present results that show improvement in accuracy and robustness using neural network-based algorithms. Additionally, talk will discuss strategies for software-hardware codesign to enable high energy efficiency without loss of accuracy.

Session Chair: Chetan Singh Thakur

Bio: Dr. Jonas is a National Board-certified Psychiatrist, who completed his MBBS and Post-graduate Diploma in Psychiatry (DPM) from Christian Medical College, Vellore, India. He further went on to obtain his MRCPsych from the Royal College of Psychiatrists, UK, and his DNB(Psych) from the National Board of Examinations, New Delhi, India. Being an experienced clinician with a keen interest in research, he transitioned to a full-time research career when he joined CBR in 2019.

Title: Neurocognitive assessment in human subjects: processes and implications (session by the clinical research team of CBR)

Abstract: Neurocognitive assessment is an integral part of the comprehensive, multimodal assessments that are central to the longitudinal cohort studies on brain aging (https://www.cbr.iisc.ac.in/research/flagship-projects/) spearheaded by the Centre for Brain Research (CBR). This session by the clinical research team of CBR will touch upon the need, rationale, and design of neuropsychological assessments. It will provide an overview of the fundamental principles of different neurological tests and domain-wise tests spanning (a) attention, visuospatial abilities, and executive function (b) learning memory, language, and social cognition (c) scoring and interpretation. It will include a demonstration of Hindi Mental State Examination (HMSE), a tool used to assess cognitive impairment.

Session Chair: Sridharan Devarajan

Bio:  Dr. Thomas Gregor Issac is currently working as an Associate Professor at the Centre for Brain Research (CBR) at Indian Institute of Science (IISC) after his DM residency (Geriatric Psychiatry) at the prestigious National Institute of Mental Health and Neurosciences (NIMHANS) Bengaluru.
He has finished his M.B B. S training from M.O.S.C Medical College (Mahatma Gandhi University) Kerala from 2003-2009 and subsequently selected for Ph.D. in Clinical Neurosciences Programme under the ICMR MD-PhD Talent search programme which he pursued from August 2011 till July 2016. His PhD dissertation focused on the role of “Renin-Angiotensin-Aldosterone System in Cognitive deterioration in patients with cerebral small vessel disease” wherein he had looked into the genetic, cognitive and neuroimaging profiles on a cohort of more than 200 patients across the spectrum of vascular cognitive impairment. Subsequently he was selected into MD Psychiatry course at NIMHANS (from July 2016- June 2019) and trained as a Junior resident in the Dept. of Psychiatry. His MD dissertation involved looking into the “Role of APO E4 allele in vascular cognitive
impairment”. He also has cleared DNB in Psychiatry. After his post-graduation in psychiatry, he got selected for DM (Geriatric psychiatry) at NIMHANS, wherein he has finished Senior Residency. His DM dissertation focussed on the “Naturalistic follow-up study on older adults with unipolar major depression after in-patient care in a tertiary centre”. He has > 50 Pubmed indexed publications

Title: Neuropsychological evaluation- Basic tenets

Abstract: The need for understanding brain functions has been explored since the begining of civilization. However, there are certain methods which can tap into the various cognitive domains and bring out the deficits if any. This is extreme;ly important as it gives an idea into the various designated functions of the brain structutes but also gives relatime insight into areas requiring cognitive rehabilitation. This talsk focusses on the role of Neuropsychological assessments as a gold standard in understanding cognitive deficits, provide a detailed neurocognitive domain based assessment and also the need for cognitve rehabilitation utilizing the plasticity properties of the Brain in various disorders especially neurodegenerative disorders.Cognitive remediation using feedback principles would form the crux of the talk along with some demonstration on how the cognitive tests are administered in real time.

Session Chair: Sridharan Devarajan

Bio: Chinna received his Ph.D. degree in neurobiology in 2016 from the University of Haifa, Israel. He carried out his Ph.D. thesis research at the University of Haifa and at the RIKEN Center for Brain Science, Japan. He uncovered the principles of associative taste aversion learning and memory. He pursued his postdoctoral research at the Picower Institute for Learning and Memory at the Massachusetts Institute of Technology, USA (2016-2020), where he continued to work as a research scientist (2021-2022). He demonstrated the neural circuit basis of cognitive dysfunctions in mouse models of Alzheimer’s disease and how brain stimulation could modify Alzheimer’s pathophysiology, learning & memory. In the summer of 2022, he joined as a faculty member at the Centre for Brain Research at the Indian Institute of Science.

Title: Non-invasive brain stimulation to impact neural circuits, sleep and cognition in Alzheimer’s disease

Abstract: A mechanistic understanding of the impact of Alzheimer’s disease (AD) on cognitive functions has been pursued across many levels of analysis; however, recent work has provided growing support for the view that it can be considered a large-scale synaptic and network disconnection disorder. For example, coherence and/or amplitude of gamma frequency neural oscillations (~30-100 Hz), which are implicated with numerous higher-order cognitive functions, are affected in various brain areas in AD. Moreover, neural circuit dysfunction plays a critical role in AD etiology; such disruptions have effects on several cell types. Thus, our working hypothesis is that correcting neuronal circuit alterations has the potential to feed back onto multiple cell populations to improve molecular and cellular pathologies. We have sought less invasive approaches for modifying neural activity and gamma oscillations. Our approach has been to harness patterned sensory and transcranial electrical stimuli, which are known to entrain network oscillations in human and animal models. I will briefly share some data which shows that non-invasive brain stimulations administered at gamma frequency improve episodic memory and gene expression programs in neurons and glial cells and impact neural oscillations and sleep.

Session Chair: Sridharan Devarajan

Bio: Chinna received his Ph.D. degree in neurobiology in 2016 from the University of Haifa, Israel. He carried out his Ph.D. thesis research at the University of Haifa and at the RIKEN Center for Brain Science, Japan. He uncovered the principles of associative taste aversion learning and memory. He pursued his postdoctoral research at the Picower Institute for Learning and Memory at the Massachusetts Institute of Technology, USA (2016-2020), where he continued to work as a research scientist (2021-2022). He demonstrated the neural circuit basis of cognitive dysfunctions in mouse models of Alzheimer’s disease and how brain stimulation could modify Alzheimer’s pathophysiology, learning & memory. In the summer of 2022, he joined as a faculty member at the Centre for Brain Research at the Indian Institute of Science.

Title: Tutorial on Non-invasive sensory brain stimulation

Session Chair: Sridharan Devarajan

Bio: Bratati is a computational biologist focusing on human genetics and genomics. Her research examines the role of genetic variants in shaping our predisposition to complex diseases, especially, neurodegeneration, particularly in the Indian population, with special emphasis on whole genome and whole exome sequencing analyses. Her group’s research is aimed to better understand the shared genetic architecture and causal relationship between cardiometabolic disorders and of psychiatric and neurodegenerative disorders, including genetic overlap between different disorders, and genetic and phenotypic heterogeneity within disorders in Indian population. She also leads the identification of genetic variants in the Indian population, from CBR, particularly in the GenomeIndia consortium .

Title: Genetic architecture of human cognition inferred from whole genome sequencing studies.

Abstract: Genetic architecture of human cognition inferred from whole genome sequencing studies Cognitive function is heritable, with metabolic risk factors known to accelerate age- associated cognitive decline. Identifying genetic underpinnings of cognition is thus crucial. Analyzing whole-exome sequencing data from >155,000 individuals for neurocognitive phenotypes at single-variant and gene level, we uncover the genetic architecture of human cognition. We found 18 independent novel loci associated with five cognitive domains while controlling for APOE isoform-carrier status and metabolic risk factors. Our novel variants are mostly in genes which could also impact cognition via their functions on synaptic plasticity and connectivity, oxidative stress, neuroinflammation. Variants in or near these identified loci show genetic links to cognitive functioning in association with APOE, Alzheimer’s disease and related dementia phenotypes and brain morphology phenotypes. These exome-wide significant variants can also substantially regulate expression of their corresponding genes in various regions of the brain. We further identify four novel pairwise interactions between exome-wide significant loci and APOE variants influencing episodic memory, and simple processing speed while accounting for serum lipid and serum glycemic traits. We obtain both main and interaction effects for APOC1 and LRP1 upon complex processing speed and visual attention. Variants in APOC1 and LRP1 are also observed to regulate their expression in basal ganglia and cerebellar hemispheres, crucial to visual attention. We also uncover variants showing evidence of pleiotropy and mediation effects through serum glucose/HDL levels affecting cognition. Our research highlights a novel set of loci that augments our understanding of the genetic underpinnings of cognition during ageing, considering co- occurring metabolic conditions that can confer genetic risk to cognitive decline in addition to APOE, which can aid in finding causal determinants of cognitive decline.

Session Chair: Sridharan Devarajan