A new arti­cle in Neu­ron from Prince­ton Uni­ver­si­ty neu­ro­sci­en­tists Flo­ra Boucha­court and Tim Buschman presents a new mod­el of work­ing mem­o­ry.

Work­ing mem­o­ry is your abil­i­ty to hold things ‘in mind.’ It acts as a work­space in which infor­ma­tion can be held, manip­u­lat­ed, and then used to guide behav­ior. In this way, it plays a crit­i­cal role in cog­ni­tion, decou­pling behav­ior from the imme­di­ate sen­so­ry world. One of the remark­able things about work­ing mem­o­ry is its flex­i­bil­i­ty — you can hold any­thing in mind.

How this flex­i­bil­i­ty is achieved has not been under­stood. In their new man­u­script, Boucha­court and Buschman present a new mod­el of work­ing mem­o­ry that cap­tures this flex­i­bil­i­ty.

The mod­el com­bines a high-dimen­sion­al ran­dom net­work with struc­tured sen­so­ry net­works to flex­i­bly main­tain any input. The untuned nature of the con­nec­tions allows the net­work to main­tain any arbi­trary input.

How­ev­er, this flex­i­bil­i­ty comes at a cost: the ran­dom con­nec­tions over­lap, lead­ing to inter­fer­ence between rep­re­sen­ta­tions and lim­it­ing the mem­o­ry capac­i­ty of the net­work. This match­es the lim­it­ed capac­i­ty of work­ing mem­o­ry in humans and sug­gests there is a trade­off between flex­i­bil­i­ty and capac­i­ty in work­ing mem­o­ry.

In addi­tion, the mod­el cap­tures sev­er­al oth­er behav­ioral and neu­ro­phys­i­o­log­i­cal char­ac­ter­is­tics of work­ing mem­o­ry.

This work pro­vides new insight into a core cog­ni­tive func­tion in humans. Ongo­ing work hopes to under­stand how these mech­a­nisms may be dis­rupt­ed in neu­ropsy­chi­atric dis­eases that dis­rupt work­ing mem­o­ry.

Sto­ry Source:

Mate­ri­als pro­vid­ed by Prince­ton Uni­ver­si­ty. Note: Con­tent may be edit­ed for style and length.

Source link