When the working memory is overloaded in humans, the synchronization between the three parts of the brain is disturbed.

Cortical architecture in accordance with the canonical model of neural microcontour in the cerebral cortex of primate. Four types of neurons (stellate neurons, superficial and deep pyramidal neurons, and inhibitory interneurons) are connected by excitatory (red) and inhibitory (black) connections. This set of neurons and compounds is motivated by anatomical and theoretical considerations in favor of the canonical model.

A person is able to simultaneously hold in the working memory a limited number of objects . The amount of working memory is directly related to cognitive ability , which decreases with neurological diseases and mental disorders. Scientists have been studying for several decades how the loading of working memory affects the processing of neural signals in the brain. They are trying to understand why the working memory is so small. And why cognitive abilities drop dramatically if you load working memory beyond the limit.

Studies on the loading of working memory and its limitations were focused on coordinating activities in the fronto-parietal network. It is known that it plays an important role in working memory . These studies predicted the limits of working memory bandwidth, measuring the level of network integration (that is, how different parts of the frontal-parietal network are connected together) and synchronizing the work of these parts of brain wave activity in the gamma range . Recent studies have revealed that visual working memory works independently for two visual half-fields (hemifields) - left (LFP) and right (RFP), and that load changes have different effects on the vibrational dynamics of different frequencies .

Now, researchers at the Massachusetts Institute of Technology's Pickauer Institute for Learning and Memory have come close to explaining why cognitive abilities decline when working memory is overloaded. They found that in this case the conjugation is violated, that is, the synchronization of the brain waves of the three key regions.
')
Scientists have developed a large-scale theoretical model of the cortical network, which consists of the prefrontal cortex (PFC), the eye fields of the frontal cortex (FEF) and the lateral intraparity region (LIP). This is an extended version of the previous model , based on predictive coding, but only here the Cross Spectral Density (CSD) responses are used to identify changes in neural connections that underlie spectral power changes at different frequencies.


Large-scale canonical model of neural microcontour

The new model allowed us to determine how the load-dependent effects affect the functional connectivity with abrupt changes in neural connectivity when the working memory capacity is exceeded and the prediction signals are destroyed.

“When you reach maximum capacity [of working memory], feedback is lost,” explains Professor Earl Miller, co-author of the scientific work. This loss of synchronization means that the three regions (PFC, FEF and LIP) can no longer interact with each other to maintain working memory: the prefrontal cortex of the PFC stops giving feedback to FEF and LIP. After a certain level of stress on the brain, low-frequency signals between the PFC, FEF, and LIP are out of sync - and the working memory is no longer functioning.

The experiment confirmed the previous discovery that the amount of working memory is independent for the left and right half-fields .

The average number of objects that can be held in the visual working memory varies among different people, but usually amounts to four objects, the professor says. The amount of working memory is usually associated with the level of intelligence.

Having made a number of scientific discoveries on the functioning of the brain and working memory of a person, the authors of the research, Earl Miller and Timothy Bushman, founded the commercial company SplitSage , which develops programs for testing human intellectual abilities and the amount of his working memory. In the future, such programs may be included in the standard test suite for knowledge workers.

The scientific article was published on March 28, 2018 in the journal Cerebral Cortex (doi: 10.1093 / cercor / bhy065).