A behavioral and ERP study of verbal mode activation as a function of time

CECILIA IMPERIOSO; SILVANO ZANUTTO; VIRGINIA JAICHENCO; ALEJANDRO WAINSELBOIM

Florianópolis

Congreso; Language and Neuroscience Conference; 2012

Universidad Federal de Santa Catarina

Two basic questions in psycholinguistic research are i) which informationis activated after a linguistic stimulus is presented, and ii) how long isthis information activated or present in working memory. In this studywe aimed to analyze these two questions regarding a specific feature ofSpanish verbs that had not been previously studied: verbal mode. To doso we carried out three lexical decision experiments with priming. Ineach of them 100 relevant prime-target pairs and 100 prime-target fillerswere presented in pseudo-random order. All prime and relevant targetwords were Spanish verbs in 3rd person plural form, while 100 targetnon-words were fillers obtained by changing one or two initial letters ofexisting Spanish verbs. 50 of the relevant prime-target pairs shared allverbal syntactic information (Same condition), while the other 50 differedonly in verbal mode between prime and target (Different condition).Primes and targets were matched in mean frequency of use and length,and shared no semantic or formal relation. Each trial consisted of a3000 ms resting period followed by prime presentation (200 ms); interstimulusinterval (Exp 1: 100; Exp 2: 250 ms; Exp 3: 450 ms) and targetpresentation (until response). Participants were instructed to respondas quickly as possible whether target stimuli were real Spanish words ornot. Electroencephalographic (EEG) recordings were gathered during thetask. EEG measures have proven to be a valuable tool to study the fastdynamic changes that underlie cognitive processing, because of their highAbstracts25temporal resolution. In addition, EEG recordings can be decomposed intooscillatory components, which allow to analyze the signal as a functionof time and frequency. Increases (event related synchronization, ERS)and decreases (event related desynchronization, ERD) in frequency bandpower after stimulus presentation are considered a measure of the degreeof synchronized neural firing over the analyzed regions. Results showedthat the percentage of correct responses increased and reaction timedecreased when prime and target words shared the same verbal mode.Furthermore, the magnitude of the priming effect was modulated by theinter-stimulus interval presented during the experiment, with a maximumvalue in Exp 2. Global mean potential value measured 50-250 ms aftertarget presentation increased in the Same compared to the Differentcondition. Similarly to behavioral results, the magnitude of the effect wasmodulated by the inter-stimulus interval used, with a maximum in Exp 2.Finally, a significant ERS of theta and alpha rhythms was found early afterprime presentation, while a significant ERD of these same frequencieswas found approximately 400 to 1000 ms after target presentation. In thisperiod a greater ERD was obtained in the Same compared to the Differentcondition. These differences are similar to those seen during visual n-backtasks in which a stronger alpha ERD is found after presentation of targetscompared to non-targets. Again, this measure was modulated as a functionof the inter-stimulus interval, with a maximum value in Exp 2. Therefore,both behavioral and EEG measures coincided to show that in the presentexperiments verbal mode seemed to attain a maximum value within 350ms after prime presentation.