Study: For cognitive training to work, it must induce neuroplasticity in brain regions that matter

Over the last sev­er­al years, cog­ni­tive train­ing has received large amounts of pub­lic inter­est and sup­port because reli­ably improv­ing cog­ni­tive per­for­mance would have wide reach­ing appli­ca­tions in clin­i­cal pop­u­la­tions, old­er adults, and the pub­lic at large. For exam­ple, cog­ni­tive train­ing could play an impor­tant role in healthy aging by delay­ing the onset of age relat­ed cog­ni­tive decline, or by blunt­ing the sever­i­ty of decline.  It could also help schiz­o­phrenic patients alle­vi­ate some of the severe­ly debil­i­tat­ing cog­ni­tive symp­toms of the dis­ease, such as deficits in the exec­u­tive func­tion­ing and work­ing memory.

As a result of this enthu­si­asm, both among the sci­en­tif­ic com­mu­ni­ty and the pub­lic, many labs across the globe have sought to demon­strate how var­i­ous cog­ni­tive train­ing par­a­digms might improve per­for­mance on untrained tasks and abil­i­ties, a phe­nom­e­non known as trans­fer. Unfor­tu­nate­ly, many cog­ni­tive train­ing stud­ies have found mixed results, with lim­it­ed or unre­li­able improve­ments over­all to untrained tasks and abil­i­ties. One impor­tant rea­son for the lack of repli­ca­tion and suc­cess in the field is an under­de­vel­oped under­stand­ing of the basic neu­ro­science of train­ing and trans­fer. The pre­dom­i­nant hypoth­e­sis of how trans­fer occurs sug­gests that train­ing will trans­fer to anoth­er task if the train­ing induces some form of plas­tic­i­ty or change in brain regions or net­works impor­tant for the untrained task. This is known as the neur­al over­lap of train­ing hypoth­e­sis (Jonides 2004; Nature Neu­ro­science; Dahlin et al. 2008; Sci­ence), and its psy­cho­log­i­cal pre­cur­sor has been around since Thorndike & Wood­worth’s sem­i­nal papers in the ear­ly 1900s. While this hypo­thet­i­cal frame­work agrees with much the lit­er­a­ture in cog­ni­tive neu­ro­science, it has not received much empir­i­cal support.

In our lat­est study, we attempt­ed to test this hypoth­e­sis by exam­in­ing whether the changes in brain activ­i­ty that occur after train­ing for 30 hours with Space Fortress, a com­plex cog­ni­tive train­ing video game, would be pre­dic­tive of indi­vid­ual dif­fer­ences in per­for­mance changes in an untrained work­ing mem­o­ry task. The Space Fortress train­ing task is quite com­plex, tar­get­ing work­ing mem­o­ry, exec­u­tive func­tion­ing, motor con­trol, and atten­tion, and cer­tain aspects of the train­ing are quite sim­i­lar to an untrained work­ing mem­o­ry task known as the Stern­berg Mem­o­ry Task. There­fore, we expect­ed to find that train­ing induced plas­tic­i­ty in regions of the brain impor­tant for work­ing mem­o­ry would pre­dict post-train­ing improve­ments in an untrained work­ing mem­o­ry task, (i.e. trans­fer of train­ing effects).

Once the study was com­plet­ed, we found evi­dence sup­port­ing this asser­tion. Specif­i­cal­ly, we found that changes in brain acti­va­tion in the tar­get areas, which we label as train­ing-induced plas­tic­i­ty, pre­dict­ed 37% of the vari­ance in the per­for­mance changes in the untrained work­ing mem­o­ry task. This sug­gests that those indi­vid­u­als that demon­strat­ed the great­est ben­e­fit from the train­ing were also those on whom the train­ing had the great­est neur­al impact. Fur­ther­more, the impor­tance of the neur­al impact was spe­cif­ic to areas of the brain already known to be involved in the cog­ni­tive con­struct that is involved in the untrained (trans­fer) task. Pre­vi­ous research in cog­ni­tive train­ing has found that untrained improve­ments simul­ta­ne­ous­ly occur with changes in brain activ­i­ty in work­ing mem­o­ry asso­ci­at­ed areas, such as the cau­date, an area known to be impor­tant for work­ing mem­o­ry and pro­ce­dur­al learn­ing (Dahlin et al. 2008), and our find­ings extend this research by demon­strat­ing that this plas­tic­i­ty actu­al­ly pre­dicts the behav­ioral changes that occur after training.

In oth­er words, cog­ni­tive train­ing can trans­fer to a real-life activ­i­ty when it changes the brain regions under­ly­ing per­for­mance in that real-life activ­i­ty.  Our find­ings offer some of the first con­fir­ma­to­ry sup­port to the neur­al over­lap of how trans­fer of train­ing occurs. Fur­ther­more, these find­ings sug­gest that future cog­ni­tive train­ing stud­ies should be designed for max­i­mum impact on brain activ­i­ty. This might be achieved, for exam­ple, by com­bin­ing cog­ni­tive train­ing with an exer­cise reg­i­men and using phys­i­cal activ­i­ty to up-reg­u­late bio­mark­ers asso­ci­at­ed with neu­ro­plas­tic­i­ty. Sim­i­lar­ly, dur­ing the cog­ni­tive train­ing ses­sions, brain stim­u­la­tion such as trans-cra­nial direct cur­rent stim­u­la­tion ‑which has also been shown to increase lev­els of sev­er­al bio­mark­ers asso­ci­at­ed with neu­ro­plas­tic­i­ty- might be applied.

For more infor­ma­tion on our study, and how it relates to the cur­rent state of the cog­ni­tive train­ing lit­er­a­ture, please see our open access paper recent­ly pub­lished in Fron­tiers in Human Neu­ro­science: Pari­etal plas­tic­i­ty after train­ing with a com­plex video game is asso­ci­at­ed with indi­vid­ual dif­fer­ences in improve­ments in an untrained work­ing mem­o­ry task

-- Aki Niko­laidis is a grad­u­ate stu­dent at the Uni­ver­si­ty of Illi­nois Urbana Cham­paign. His work focus­es on under­stand­ing how brain plas­tic­i­ty con­tributes to learn­ing and enhance­ment of cog­ni­tive abilities.

Relat­ed articles:

• Does brain train­ing work? Yes, if …
  
• How to eval­u­ate brain train­ing programs