skip to primary navigation skip to content

CBSU bibliography search

To request a reprint of a CBSU publication, please click here to send us an email (reprints may not be available for all publications)

Memory traces for words in the brain are indexed by the mismatch negativity (MMN) responses: fMRI evidence
Neuroimage 31, S116
Year of publication:
CBU number:
Introduction We have earlier demonstrated that the mismatch negativity (MMN) response is enhanced for word stimuli as compared with acoustically matched phonologically legal meaningless pseudowords (1-4). This enhancement of the word-elicited MMN, or lexical Representational Negativity (RN), is taken as a neurophysiological signature of word-specific memory networks/cell assemblies activated in the human brain by acoustic input in an automatic and attention-independent fashion. This phenomenon has been replicated by a number of studies using different techniques and various languages (for review, see 5). The question however still remains whether it can be documented using haemodynamic brain imaging, which could provide a more exact information on the location of the word-specific activity than the currently available electrophysiological data. This has been the aim of the present study. Methods In two conditions, word stimuli were randomly presented (P=12.5%) as deviants among acoustically matched pseudowords in trains of stimuli which were played in silence and followed by the fMRI scanning sequence during which no stimuli were presented (so-called 'sparse imaging'). In two other conditions (the order being counterbalanced), the reverse was used and infrequent pseudowords were interspersed with standard word stimuli. The subjects were instructed to ignore the auditory stimulation and concentrate on watching a self-selected video. Echo planar images were obtained using a 3T Bruker MR system (TR 3.02s, TE 115ms, flip angle 90º, 21 4mm-thick slices covering the whole brain, interslice distance 1mm, in-plane resolution 1.6x1.6mm). Imaging data were processed using SPM2000. The contrasts of interest included all auditory events vs. null events, deviants vs. standards (MMN), and, most crucially, word-evoked MMN vs. pseudoword MMN. Results Random-effect analysis indicated strong auditory activation in both auditory cortices. However, the pattern of this activation was clearly distinct between the word and pseudoword blocks. Whereas MMN was found as being significant in the word conditions, it did not reach the significance threshold for the pseudoword conditions. This effect was more pronounced in the left than in the right hemisphere. To further analyse this, we extracted average parameter estimates in voxels within 15mm sphere around the centre of Heschl's gyrus in each subject's left and right hemispheres, and submitted these to ANOVA. This confirmed differential activation and produced significant hemisphere by stimulus type interaction indicating stronger BOLD MMN responses to words than pseudowords in the left but not in the right hemisphere's auditory cortex. Additionally, results from scans obtained at different delays after the deviant presentation in the train indicated possibly diverging temporal dynamics of word- and pseudoword-related activations. Conclusions Our results confirm the existence of the automatic pre-attentive word-related MMN enhancement, which can reflect the presence of a memory trace for a spoken word, as previously suggested. For the first time, this enhancement is documented using fMRI, which indicated strong involvement of auditory structures in the vicinity of left-hemisphere's Heschl's gyrus in sustaining such word-specific cortical networks and suggested a specific temporal patterns of their activation. References 1.Pulvermuller F. et al, NeuroImage, 2001 2.Pulvermuller F. et al, Psychophysiology, 2004 3.Shtyrov Y. et al. Neuroreport, 2002 4.Shtyrov et al. NeuroImage 2005 5.Pulvermuller F., Shtyrov Y. Progress in Neurobiology, 2006