The idea of reverse-engineering the brain is a stimulating challenge with a plethora of potential outcomes and applications. However, much of contemporary brain research is like the proverbial blind men and the elephant. Each blind man senses a different part of the animal and describes their perspective, but as a group they fail to develop a comprehensive understanding. Modern neuroscience is plagued by these unconnected views. Specifically, many neuroscientists study brain function using only a small fraction of the available data collection tools. Some focus on coarse-scale EEG signals that can be recorded from non-invasive electrodes placed outside the skull. Others study a single cell type using sophisticated neural probes or invasive optical imaging techniques with a portion of the skull removed. Still others will focus on vasculature in the brain and how blood oxygenation levels reflect neural activity. Unfortunately, there is a significant gap in our understanding of brain function between these various scales.
This workshop is aimed at addressing the following Grand Challenges of neuroengineering:
- Develop a comprehensive understanding of the role of different cell types in brain function. Much of neuroscience has historically focused on the anatomy and function of neurons in isolation, but emerging research suggests other types of cells, such as glial cells, and neurovascular coupling play a critical role in coordinating and modulating neural activity.
- Unify our understanding of neural activity across multiple scales. Technologies such as fMRI, fNIR, EEG, MEG, novel wearable biomonitors, and neural probes allow different aspects of the brain’s response to external stimuli to be quantified, at different spatial and temporal resolutions and different costs and levels of invasiveness. Typically only one type of measurement is used in any given neuroscience study, limiting the scope of any outcomes. Engineering instruments for data fusion, signal classification, and multi-modal analysis present enormous opportunities for addressing this significant limitation.
- Investigate the fundamental relationship between brain form and function. Many higher-level brain functions, such as vision and memory, require large numbers of neurons to act in concert. Understanding the connectivity patterns of neurons across various scales may lead to new insights into the organizing principles of the brain, helping investigators:
- better infer specific brain functions from observations of connectivity,
- better understand the role of connectivity patterns in variations in cognitive abilities across populations, and
- perform “mind reading” — predicting tasks or stimuli based on brain activity
- Design new devices, circuits, and systems capable of simulating brain activity and performing human-level pattern recognition. A critical test of science’s understanding of the brain is its ability to develop a synthetic system at a similar scale and level of performance. Accomplishing this task not only requires a fundamental understanding of brain function across multiple scales, but also the development of novel device technologies. Furthermore, current machine learning systems are incapable of performing as well humans on a diverse set of tasks; for instance, visual scene understanding and natural language processing are much challenging for machines than people expect in small special cases. Novel hardware architectures and integrated software inspired by natural neural systems may allow us to overcome this hurdle without supercomputers.