Dissertation Plasticity in Eeg Oscillations Associated with Auditory Verbal Learning

P L A S T IC IT Y IN E E G O SC IL L A T IO N S A SSO C IA T E D W ITH A U D IT O R Y V E R B A L LEA R N IN G O n a frequent basis, hum ans need to vocally learn and rem em ber a lis t o f un re la ted item s. A dvances in cognitive neuroscience have begun to iden tify the bra in reg ions involved. H ow ever, the m echanism s b y w h ich those reg ions in terac t during learn ing rem ain elusive. T here is grow ing support fo r the p roposition th a t the oscilla tions w ith in and am ong these reg ions prov ide a substrate for the ir in teraction . T h is proposition is investigated in the p resen t study by evaluating changes in b ra in osc illa tions during verbal learning. P rev ious research in th is dom ain has provided on ly lim ited clues about the in fluence o f eco logically sign ifican t factors such as repetition and m nem onics on learn ing perform ance and b ra in dynam ics. T he p resen t study evaluates independen t com ponents analysis, pow er spectral analysis, and coherence o f 32-channel e lectroencephalogram reco rd ed w hile subjects learned a lis t o f un re la ted nouns. T he learn ing task included rep e titio n and e ith er conventional spoken learn ing or learn ing w ith a m usical m nem onic. The resu lts show tha t as subjects m ake the transition from repetition to learn ing , th e ir alphai frequency b and ac tiv ity undergoes a state transition from synchronized to desynchronized oscillations over rig h t poste rio r cortex. A sim ilar state transition is observed w hen learning includes a m usical m nem onic, bu t its topographic d istribu tion in the rig h t hem isphere is reversed and relative desynchrony occurs over righ t p refrontal cortex. The resu lts suggest tha t verbal learning, even in the con tex t o f repetition , is associated w ith m odula tion o f b ra in oscilla tions and that an anatom ically d istinc t ne tw o rk is recru ited w hen learn ing includes a m usical m nem onic. T he study has im plications fo r the basic cognitive neuroscience o f learn ing , c lin ica l rehabilita tive applications using learn ing m nem onics, and the arch itectures o f b io logica lly -p lausib le m achine learning algorithm s. D avid A . P eterson D epartm en t o f C om puter Science C olorado State U niversity F ort C ollins, C O 80523 Sum m er 2007 iii Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. ACKNOW LEDGEM ENTS This research would not have come to fruition without the help o f many other people. My advisor Chuck Anderson has provided undying support for the uniquely interdisciplinary path I have followed in pursuit o f the PhD. I am fortunate to have worked with him on such an exciting array o f research projects over the years. His patience and guidance have been steadfast and are greatly appreciated. Bruce Draper has been an excellent source o f constructive input at multiple points in my PhD path, from research on modeling pilot training early in my program to some o f the gross anatomical underpinnings o f the research presented herein. Jim Bieman has on numerous occasions helped pull me out o f the trees to see the forest, w hether it was a hypothetical framework for this research or the career implications o f directing my computer science research into the biosciences. Darrell W hitley provided a valuable critique on the direction o f an early instantiation o f my proposed research. I am particularly grateful for the opportunity to have worked with M ichael Thaut. His views on basic operating principles o f the brain deeply resonate with my own. He has instilled in me an appreciation for how music serves as a conduit to understanding the brain. He has also helped show me the synergistic benefits o f working with clinical populations for both basic scientific inquiry and for progress toward therapeutic applications. Edward DeLosh has willingly supported my delving into the cognitive scientific realm, from the day I first inquired about the Psychology department's EEG facilities, through our early attempt at writing an NIH grant proposal, to his input on interpreting the results o f this particular research. At an institutional level, I owe a special thanks to the multiple departments at Colorado State that have at various times provided me with the facilities and equipment to conduct this research. The Computer Science department gave me an excellent foundation in my graduate studies and I have always found it to be a friendly, supportive environment. Several other units,