engleski

Recent advances in the neuroscience of learning and memory

Learning, memory, and cognition are some of the most fundamental mental processes. Various approaches have been used to understand mechanisms underlying these processes, including artificial intelligence (AI), with the main goal to clarify the role of synaptic plasticity. Recent technological advances such as optogenetic stimulation, in which light-activated proteins like channelrhodopsin are expressed in specific neurons and used to modulate cells’ activity by light, transformed the research on neural circuits by providing tools to analyze old questions. Traditionally, it was thought that outstanding talents, such as extraordinary memory, could be found only in intellectually exceptional individuals. However, some people with savant syndrome exhibit remarkable abilities in some domains of knowledge. Savant syndrome is associated with a left-hemisphere dysfunction. Unlike the healthy mind, which has built-in expectations about the world, the mind of a savant syndrome person must simplify the world by adopting strict routines. As a consequence, it seems that savants have privileged access to low- level, unprocessed information, before it is "packaged" by meaningful processing of the left hemisphere. Typically, they concentrate more on the parts than on the whole, which is a characteristic of so-called autistic geniuses. This association was raised in a recent movie "Magnus" (2016), chronicling the life of world chess champion Magnus Carlsen, who became a grandmaster at age 13. The enhanced memory capacity can be also seen in some patients with frontotemporal dementia with predominant left hemisphere involvement. Similarly, it can also be induced in normal people by creating a virtual injury of the left temporal lobe using repetitive transcranial magnetic stimulation. In addition to advancements in human research, a breakthrough achievement has been accomplished in the field of AI, as the AlphaZero program achieved superhuman performance in board games without any domain knowledge except the game rules. By using only an “empty” artificial neural network and a general reinforcement learning algorithm from self-play it defeated the world’s best chess program made in cooperation with grandmasters (winning 28, drawing 72, and losing none in a victory over Stockfish). Although we are still far from understanding the human brain and its synaptic plasticity, it is believed that machine deep learning together with advancements in optogenetics and non-invasive brain stimulation will help us to not only acquire a better understanding of the mechanistic underpinnings of various neurological and psychiatric conditions, but also to facilitate improvements in treating difficulties in learning, memory, and cognition observed in these disorders.

learning ; memory ; cognition ; artificial intelligence ; synaptic plasticity ; optogenetics ; non-invasive brain stimulation

nije evidentirano