Scientific American article about SC2+Science - Page 5

On December 02 2011 10:01 FallDownMarigold wrote:
Very interesting read, thank you for the heads up about the article. The fact that you are a researcher who likes SC2 tempts me to ask you something, so I'd love to see you thoughts

I'm wondering, what do you think about the idea of removing the motor requirement for SC2? In other words, what do you think about direct control over SC2 units via neuromotor prostheses (NMP)? The reason I'm asking is because one of the important things the article highlights is the motor aspect of SC2's difficulty. It's not like chess, where motor skills aren't really that relevant. Sure, you need basic motor connectivity to play chess, but I would argue that you don't need the refined, pruned, and much more specified corticospinal motor neuronal plasticity that you do need at high level SC2.

So, on to my rationale: It has been demonstrated in mouse, macaque, and human that neuronal ensembles (activation patterns of many neurons firing in concert to some end) unique to specific, intentional motor commands exist at the primary cortex (M1) even when corresponding limbs are no longer present or innervated. Recently, researchers have developed novel strategies combining brain-computer interface technology with algorithms to recognize and record unique, intentional neuronal ensembles at M1 corresponding to unique, intentional movements. Test patients have had electrodes implanted at M1 capable of recording and analyzing cortical motor activity. Using advances in brain-machine interface, NMPs have been developed whereby patients fitted with M1 electrodes can transfer intentional thoughts directly to a machine in order to achieve movement. For example, a patient can move a cursor around a screen based on neuronal firing patterns at the patient's M1 cortical region. First researchers determine which patterns correspond to which commands. Researchers then input these instructions, and then the NMP can match neuronal patterns to "learned" commands, thereby enabling direct neuronal control over machine-assisted movement. Therefore, it would be reasonable to apply this concept of NMP-assisted control of movement to SC2. Specifically, it would be reasonable to develop an NMP that can record neuronal patterns unique to commands in SC2, such as "stutter step marines", or "move units X distance". In theory, any number of complex commands could be matched by statistical analysis to a unique neuronal firing pattern. These patterns could be read by the NMP directly off the player's brain, enabling the player to issue commands to SC2 unfettered by motor-requirements.

So do you think it would be interesting to study SC2 if APM was no issue? It might be a novel way to examine players executing strategy at the highest possible level where literally, there are no mechanical barriers. So long as the player knows what must be done, the player can think, and the task will be executed (provided the NMP is made very well)