\n<\/p>\n
AI (Artificial Intelligence) concept. 3D rendering.<\/span><\/p>\n getty<\/small><\/div>\n<\/div>\n<\/figure>\n On a very basic level, many of us understand that artificial intelligence is helping scientists to better understand the workings of the human brain itself \u2013 but how?<\/p>\n An interesting insight comes from a presentation by Allison Hamilos called: The Neuroscience of Spontaneity & Decision-Making. <\/strong>Hamilos has a bachelor’s degree in chemistry and biology from right here at MIT, and a PhD from Harvard. She\u2019s also a member of the Harvard-MIT Health Science and Technology Program.<\/p>\n In her presentation, which she gave at a recent Science Dinner here in Boston, Hamilos presents some striking findings about human movement and how it happens: how the brain gives us the impulses that make us move around.<\/p>\n Hamilos starts by explaining the workings of \u201cmotor neurons\u201d and how many of these, working together, help to generate movement triggers.<\/p>\n \u201cThousands of motor neurons suddenly start firing together in concert, and this avalanche of neural activity sends signals down through the spinal cord to cause muscle fibers to contract,\u201d she said. \u201cThis is why we move.\u201d<\/p>\n But the work going on now is to figure out how those little avalanches get spawned.<\/p>\n In search of this answer, Hamilos makes a very critical distinction, between movement that is immediately reactive, sometimes almost involuntary, and movement that is elective, which can seem so capricious that it\u2019s hard to analyze exactly why it happens at a particular instant.<\/p>\n \u201cNot all movements result from abrupt sensory events,\u201d Hamilos notes, and it\u2019s those movements that really incite the most curiosity.<\/p>\n By way of at least partial explanation, Hamilos talks about Parkinson\u2019s disease, where science has found that the underlying \u201cmotivators\u201d are impaired or inhibited in some way. Movement, she notes, is slower, on the whole. There\u2019s also the concept of \u201cparadoxical kinesia,\u201d where a subject might display a discrepancy that looks like this: he or she might be able to respond quickly to reactive stimulus, like a football flying at someone\u2019s head, but still be inhibited on those elective motions, like picking up a book, or rising from a chair, absent any external stimulus.<\/p>\n<\/section>\n Hamilos explains that scientists have identified a common culprit, dopamine, as active in triggering motor neuron activity. But the dopamine, she adds, is probabilistic in effect, not deterministic. So humans still have free will.<\/p>\n It turns out a lot of research has been done on mice, and Hamilos goes into great detail. Successful observations on interrupted movement are one part of the equation.<\/p>\n<\/section>\n Later in the talk, I picked out a list of three components that Hamilos mentions in motion analysis: what movements to make and when, what option to choose (of multiple options,) and whether you perceive something.<\/p>\n \u201cWe know we don\u2019t want to do the same thing all the time,\u201d In fact sometimes the ability to make random choices can mean life or death,\u201d she says, giving the example of a mouse trying to evade an owl. \u201cIf your movements \u2026 are predictable, you\u2019re dead meat. Instead, the mouse\u2019s best chance is to choose unexpected movements, and unleash them at unpredictable times, by which she will hopefully buy herself a reprieve.\u201d<\/p>\n So that starts to explain the issue of what\u2019s so great about probabilistic motion.<\/p>\n<\/section>\n Of course, the science of motion is much more complex in humans, with so many sophisticated, intangible and indirect reasons to make a move, of any part of the body, in any given direction.<\/p>\n Hamilos takes us through more of the data, for example, illustrating connections between delayed motor activity and \u201cbradyphrenia,\u201d a slowing down of thinking, and the phenomenon of perseveration, where one might choose the same option over, and over, and over.<\/p>\n \u201cEven emotionally, many patients seem stuck and apathetic, without the full range of affect. It was as if these patients suffered from a lack of spontaneity of any kind, difficulty generating any new thought, or action, or feeling, without some kind of external prompting.\u201d<\/p>\n She then highlights the opposite problem, too much random motion or action, citing examples like Tourette\u2019s, noting:<\/p>\n \u201cThey are stuck making these irrepressible, apparently random body movements.\u201d<\/p>\n In response, Hamilos brings up a number of other ideas: behavioral stochasticity, for one, and a \u201cshared dopamine circuit mechanism\u201d that could unite a lot of disparate medical research under one umbrella.<\/p>\n \u201cI propose that a fundamental feature of this circuit is the ability to coordinate self-generated neural activity,\u201d she says.<\/p>\n There you have it \u2013 a look into what makes us move, as humans. It\u2019s not a small detail in behavioral research. For those who suffer from either depressive conditions like Parkinson\u2019s, or manic ones like Tourette\u2019s, it\u2019s everything. And for the rest of us, there are fundamental questions in play. Why do we do what we do? To the extent that AI can start to answer this, it really does pave the way for a future that is very different from what we have known.<\/p>\n<\/section>\n<\/div>\nThe Landscape<\/h2>\n
Experimentation with Dopamine<\/h2>\n
Three Ingredients<\/h2>\n
More Evaluation of Humans<\/h2>\n