Since the early days of research into language and the brain, word meaning was assumed to be processed in specific brain regions, which most modern neuroscientists localize to the left temporal lobe. Here we use event-related fMRI to show that action words referring to face, arm, or leg actions (e.g., to lick, pick, or kick), when presented in a passive reading task, differentially activated areas along the motor strip that either were directly adjacent to or overlapped with areas activated by actual movement of the tongue, fingers, or feet. These results demonstrate that the referential meaning of action words has a correlate in the somatotopic activation of motor and premotor cortex. This rules out a unified "meaning center" in the human brain and supports a dynamic view according to which words are processed by distributed neuronal assemblies with cortical topographies that reflect word semantics.

No abstract

The processing of spoken language has been attributed to areas in the superior temporal lobe, where speech stimuli elicit the greatest activation. However, neurobiological and psycholinguistic models have long postulated that knowledge about the articulatory features of individual phonemes has an important role in their perception and in speech comprehension. To probe the possible involvement of specific motor circuits in the speech-perception process, we used event-related functional MRI and presented experimental subjects with spoken syllables, including [p] and [t] sounds, which are produced by movements of the lips or tongue, respectively. Physically similar nonlinguistic signal-correlated noise patterns were used as control stimuli. In localizer experiments, subjects had to silently articulate the same syllables and, in a second task, move their lips or tongue. Speech perception most strongly activated superior temporal cortex. Crucially, however, distinct motor regions in the precentral gyrus sparked by articulatory movements of the lips and tongue were also differentially activated in a somatotopic manner when subjects listened to the lip-or tongue-related phonemes. This sound-related somatotopic activation in precentral gyrus shows that, during speech perception, specific motor circuits are recruited that reflect phonetic distinctive features of the speech sounds encountered, thus providing direct neuroimaging support for specific links between the phonological mechanisms for speech perception and production.cell assembly ͉ functional MRI ͉ perception-action cycle ͉ mirror neurons ͉ phonetic distinctive featue N eurological theories of language have a long-standing tradition of distinguishing specialized modular centers for speech perception and speech production in left superior temporal and inferior frontal lobes, respectively (1-3). Such separate speech-production and -perception modules are consistent with a number of neuroimaging studies, especially the observations that frontal circuits become most strongly active during speech production and that speech input primarily activates the left superior temporal gyrus and sulcus (4-6). Superior temporal speech-perception mechanisms in humans may be situated in areas homologous to the auditory belt and parabelt areas in monkeys (5,7,8). In macaca, this region includes neurons specialized for species-specific calls (9, 10). Therefore, it appeared to be reasonable to postulate a speech-perception module confined to temporal cortex specifically processing acoustic information that is immanent to speech.In contrast to this view, neurobiological models have long claimed that speech perception is connected to production mechanisms (11-16). Similar views have been proposed in psycholinguistics. For example, the direct realist theory of speech perception (17, 18) postulates a link between motor and perceptual representations of speech. According to the motor theory of Liberman et al. (19,20), speech perception requires access to phoneme representations that are c...

Transcranial magnetic stimulation (TMS) was applied to motor areas in the left language-dominant hemisphere while right-handed human subjects made lexical decisions on words related to actions. Response times to words referring to leg actions (e.g. kick) were compared with those to words referring to movements involving the arms and hands (e.g. pick). TMS of hand and leg areas influenced the processing of arm and leg words differentially, as documented by a significant interaction of the factors Stimulation site and Word category. Arm area TMS led to faster arm than leg word responses and the reverse effect, faster lexical decisions on leg than arm words, was present when TMS was applied to leg areas. TMS-related differences between word categories were not seen in control conditions, when TMS was applied to hand and leg areas in the right hemisphere and during sham stimulation. Our results show that the left hemispheric cortical systems for language and action are linked to each other in a category-specific manner and that activation in motor and premotor areas can influence the processing of specific kinds of words semantically related to arm or leg actions. By demonstrating specific functional links between action and language systems during lexical processing, these results call into question modular theories of language and motor functions and provide evidence that the two systems interact in the processing of meaningful information about language and action.

The human visual system is an intricate network of brain regions that enables us to recognize the world around us. Despite its abundant lateral and feedback connections, object processing is commonly viewed and studied as a feedforward process. Here, we measure and model the rapid representational dynamics across multiple stages of the human ventral stream using time-resolved brain imaging and deep learning. We observe substantial representational transformations during the first 300 ms of processing within and across ventral-stream regions. Categorical divisions emerge in sequence, cascading forward and in reverse across regions, and Granger causality analysis suggests bidirectional information flow between regions. Finally, recurrent deep neural network models clearly outperform parameter-matched feedforward models in terms of their ability to capture the multi-region cortical dynamics. Targeted virtual cooling experiments on the recurrent deep network models further substantiate the importance of their lateral and top-down connections. These results establish that recurrent models are required to understand information processing in the human ventral stream. deep neural networks, virtual cooling Significance StatementUnderstanding the computational principles that underlie human vision is a key challenge for neuroscience and could help improve machine vision. Feedforward neural network models process their input through a deep cascade of computations. These models can recognize objects in images and explain aspects of human rapid recognition. However, the human brain contains recurrent connections within and between stages of the cascade, which are missing from the models that dominate both engineering and neuroscience. Here we measure and model the dynamics of human brain activity during visual perception. We compare feedforward and recurrent neural network models and find that only recurrent models can account for the dynamic transformations of representations among multiple regions of visual cortex. now published in PNAS: https://www.pnas.org/content/early/

Single words and sentences referring to bodily actions activate the motor cortex. However, this semantic grounding of concrete language does not address the critical question whether the sensory–motor system contributes to the processing of abstract meaning and thought. We examined functional magnetic resonance imaging activation to idioms and literal sentences including arm- and leg-related action words. A common left fronto-temporal network was engaged in sentence reading, with idioms yielding relatively stronger activity in (pre)frontal and middle temporal cortex. Crucially, somatotopic activation along the motor strip, in central and precentral cortex, was elicited by idiomatic and literal sentences, reflecting the body part reference of the words embedded in the sentences. Semantic somatotopy was most pronounced after sentence ending, thus reflecting sentence-level processing rather than that of single words. These results indicate that semantic representations grounded in the sensory–motor system play a role in the composition of sentence-level meaning, even in the case of idioms.

No abstract

Noise-normalization has been shown to partly compensate for the localization bias towards superficial sources in minimum norm estimation. However, it has been argued that in order to make inferences for the case of multiple sources, localization properties alone are insufficient. Instead, multiple measures of resolution should be applied to both point-spread and cross-talk functions (PSFs and CTFs). Here, we demonstrate that noise-normalization affects the shapes of PSFs, but not of CTFs. We evaluated PSFs and CTFs for the MNE, dSPM and sLORETA inverse operators, on the metrics dipole localization error (DLE), spatial dispersion (SD) and overall amplitude (OA). We used 306-channel MEG configurations obtained from 17 subjects in a real experiment, including individual noise covariance matrices and head geometries. We confirmed that for PSFs DLE improved after noise normalization, and is zero for sLORETA. However, SD was generally lower for the unnormalized MNE. OA distributions were similar for all three methods, indicating that all three methods may greatly underestimate some sources relative to others. The reliability of differences between methods across subjects was demonstrated using distributions of standard deviations and p-values from paired t-tests. As predicted, the shapes of CTFs were the same for all methods, reflecting the general resolution limits of the inverse problem. This means that noise-normalization is of no consequence where linear estimation procedures are used as “spatial filters.” While low DLE is advantageous for the localization of a single source, or possibly a few spatially distinct sources, the benefit for the case of complex source distributions is not obvious. We suggest that software packages for source estimation should include comprehensive tools for evaluating the performance of different methods.