A maze experiment with rats is one of the most classic brain science research methods. However, a new test prepared by a collaborative team of international researchers shows how such experiments push the boundaries of neuroscience and technology alike.
Using electrical activity, which is recorded from key brain areas, scientists have read the thoughts of rats after they run mazes or when they dream of doing so. This is a critical step in preserving memory. Researchers involved in developing mind-reading methods can now read out those signals with an almost instantaneous sense of statistical significance and high levels of accuracy. Statistical significance is a way of determining by an analyst the results in the data. They cannot be explained by chance alone.
One such team was able to achieve unprecedented speeds of decoding and analysis by implementing their neural decoding software on a graphics processing unit (GPU) chip, the same kind of hardware favored by video gamers. According to the study, the GPU-based system was 20-50 times faster than using conventional multi-core CPU chips.
Furthermore, it showed that the system remained accurate and fast even when handling more than a thousand input channels. The inputs being the rat's brain activity. This approach to real-time decoding is important because it allows for new high-density recording devices. Devices such as high-density electroencephalography (HD-EEG) which can be used to map brain areas involved in movement, sensation, language, emotion, and cognition. Another example being the Neuropixels probe which promises to measure cellular brain activity more precisely and at larger scales.
Comments