240, p < .0001). As can be seen in Appendix B, there were no main effects (or indeed interactions) of lexical category Pexidartinib or semantic-abstractness on psycholinguistic properties of stimuli. This being the case, we were confident that brain activation

in contrasts focusing on lexical category and semantic-abstractness were free of ulterior confounding effects. The experimental word categories were dispersed among 200 filler words during presentation, with which they were matched in length (F(1, 359) = 1.006, p > .436), bigram (F(1, 359) = 1.679, p > .084) and trigram frequency (F(1, 359) = .868, p > .560). 120 hash marks, matched to word stimuli in length, acted as a low level visual baseline in contrasts. Adopting a paradigm previously employed for investigating lexicosemantic Bortezomib processing (e.g., Hauk et al. 2004; for review, see Pulvermüller et al. 2009), words written in lowercase letters were presented tachistoscopically while haemodynamic responses were recorded using event-related fMRI. This passive reading paradigm was chosen to be unbiased

towards semantic or grammatical processing. Despite no overt instructions for semantic processing, it is reliably known to evoke early differential activations that reflect a word’s semantic category (see Hauk et al., 2008, for review), strongly implying that reading automatically evokes semantic processing of word stimuli in typical participants. Subjects were instructed to attend to and carefully read all stimulus words silently, without moving their lips or tongue. The passive reading task was delivered in three blocks of approximately 7 min each. A short presentation time of 150 ms ensured that saccades were discouraged and that participants had to continuously attend to the screen in order to perform the task. A central fixation cross was displayed between stimuli for an average 2350 ms, with a jitter of ±250 ms, resulting in SOAs

between 2250 and 2750 ms (average 2500 ms). The order of stimuli was pseudo-randomised (restriction: not more than two items of the same category in direct succession) with two lists, counter-balanced across subjects. Following the scan, our participants were requested to complete a short unheralded word recognition test outside the scanner. In the recognition test, they were presented SPTLC1 with a list of experimental stimuli and novel words and had to rate each word on a scale from 1 to 7, indicating how certain they were that a given item had appeared in the fMRI experiment. For evaluation, ratings were converted into percentage correct/incorrect responses. The test contained a combination of 50 experimental and 25 novel distracter words, and above chance performance was thus taken to confirm that subjects had engaged with the task. A Siemens 3T Tim Trio (Siemens, Erlangen, Germany) with a head coil attached was employed during data collection.