modelling the nervous system
Psychology provides us with an overview of what we are doing, while neuroscience informs us about the workings of the machinery behind our behaviour, but what is the use of computational modelling?
The neuroscientist David Marr proposed that information processing systems like the brain can be better understood by thinking about their function in different ways at different levels. The highest identifies the problem that is being solved, and the input and output to the system (sensory data-->behaviour). The next level down is concerned with determining the process needed to get from input to output (an algorithm). The lowest level looks at how the process is implemented physically (in neurons, for example).
Psychology and neuroscience focus on the highest and lowest levels respectively. However, there is still a large gap in our knowledge about how neurons produce movement, not to mention emotion and consciousness. Computational modelling provides an essential tool for inspection in this gap. When the theory of the configuration of a network of neurons in the brain has been created, a computational model of this neural network allows us to see if the system produces the expected behaviour. It can then easily be adjusted and adapted to answer research questions and help us pose new questions.
There are, of course, a number of sub-levels filling the gap between psychology an neuroscience, and depending where you look or who you talk to, these may be referred to as: neural circuits, systems neuroscience, theoretical neuroscience, computational neuroscience, cognitive neuroscience, cognitive psychology, connectomics, and more.
I'm interested in the question of how we (animals) have reached our current condition, particularly with respect to the nervous system. How do collections of connected neurons make us who we are, from our simple reflex arcs to music appreciation and other high level behaviours? What can we take from what we learn from the nervous system and nature in general to help us create practical and efficient applications?
I think this second question is of particular importance because as the human population of the world grows, so does the need to achieve maximally efficient performance in many different tasks (power generation, food production, waste recycling, factory-line assembly, many more). Each human still is still endowed with the evolutionary animal instinct to exploit its environment as much as possible while it is here (whether it's making money or just creating a comfortable home), so the idea of cutting back on resource consumption is impractical. The only way forward is to work out intelligent production and recycling techniques.
I did my undergrad in cognitive science and Masters' in cybernetics, starting out by trying to find answers through philosophy, then into psychology, neuroscience, and finally towards computational neuroscience.
- Neuroscience Online, the Free Neuroscience Electronic Textbook. Great for learning the basics of neuroscience. Some chapters have interactive/animated illustrations.
- A series of neuroscience videos from the Allen Institute of Brain Science.
- Webvision is an up to date and detailed book covering all areas of visual processing in the nervous system.
- Eye, Brain & Vision online book by one of the most famous vision scientists, David Hubel.
- Explorations in Parallel Distributed Processing: A Handbook of Models,Programs, and Exercises. This is an online book written by one of the fathers of neural network modelling; James McClelland. The models are made in Matlab.
- The practice of theoretical neuroscience: an editorial from Nature Neuroscience that addresses the roles of theoretical and experimental neuroscience.
- Computational Neuroscience: about the relationship between computational (theoretical) neuroscience and cognitive science.
This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivs 3.0 Unported License.