New Paper by Lim, Wöst­mann, & Obleser in Jour­nal of Neuroscience

Selec­tive atten­tion to a task-rel­e­vant stim­u­lus facil­i­tates encod­ing of that stim­u­lus into a work­ing mem­o­ry rep­re­sen­ta­tion. It is less clear whether selec­tive atten­tion also improves the pre­ci­sion of a stim­u­lus already rep­re­sent­ed in mem­o­ry. Here, we inves­ti­gate the behav­ioral and neur­al dynam­ics of selec­tive atten­tion to rep­re­sen­ta­tions in audi­to­ry work­ing mem­o­ry (i.e., audi­to­ry objects) using psy­chophys­i­cal mod­el­ing and mod­el-based analy­sis of elec­troen­cephalo­graph­ic sig­nals. Human lis­ten­ers per­formed a syl­la­ble pitch dis­crim­i­na­tion task where two syl­la­bles served as to-be-encod­ed audi­to­ry objects. Valid (vs neu­tral) retroac­tive cues were pre­sent­ed dur­ing reten­tion to allow lis­ten­ers to selec­tive­ly attend to the to-be-probed audi­to­ry object in mem­o­ry. Behav­ioral­ly, lis­ten­ers rep­re­sent­ed audi­to­ry objects in mem­o­ry more pre­cise­ly (expressed by steep­er slopes of a psy­cho­me­t­ric curve) and made faster per­cep­tu­al deci­sions when valid com­pared to neu­tral retrocues were pre­sent­ed. Neu­ral­ly, valid com­pared to neu­tral retrocues elicit­ed a larg­er fron­to­cen­tral sus­tained neg­a­tiv­i­ty in the evoked poten­tial as well as enhanced pari­etal alpha/low-beta oscil­la­to­ry pow­er (9–18 Hz) dur­ing mem­o­ry reten­tion. Crit­i­cal­ly, indi­vid­ual mag­ni­tudes of alpha oscil­la­to­ry pow­er (7–11 Hz) mod­u­la­tion pre­dict­ed the degree to which valid retrocues ben­e­fit­ted indi­vid­u­als’ behav­ior. Our results indi­cate that selec­tive atten­tion to a spe­cif­ic object in audi­to­ry mem­o­ry does ben­e­fit human per­for­mance not by sim­ply reduc­ing mem­o­ry load, but by active­ly engag­ing com­ple­men­tary neur­al resources to sharp­en the pre­ci­sion of the task-rel­e­vant object in memory.