Action-related semantic content and negation polarity modulate motor areas during sentence reading: An event-related desynchronization study

Abstract

Our study evaluated motor cortex involvement during silent reading of sentences referring to hand actions. We aimed at defining whether sentential polarity (affirmative vs. negative) would modulate motor cortex activation using the event-related desynchronization (ERD) analysis of the mu rhythm. Eleven healthy volunteers performed a reading task involving 160 sentences (80 affirmative: 40 hand-related, 40 abstract; 80 negative: 40 hand-related, 40 abstract). After reading each sentence, subjects had to decide whether the verb was high or low frequency in Italian. Electroencephalographic (EEG) activity was recorded with 32 surface electrodes and mu ERD analyses were performed for each subject. Hand-action related sentences induced a greater mu ERD over the left premotor and motor hand areas compared to abstract sentences. Mu ERD was greater and temporally delayed when the hand-related verbs were presented in the negative versus affirmative form. As predicted by the “embodied semantic” theory of language understanding, motor areas were activated during sentences referring to hand actions. In addition, motor cortex activation was larger for negative than affirmative motor sentences, a finding compatible with the hypothesis that comprehension is more demanding in the specific case of motor content negation.

Introduction

The neural basis of language is a central topic of cognitive neuroscience. In particular, knowing how the brain represents words' meaning is a fascinating question mark for both philosophers and scientists.

According to classical cognitive theories, concepts are represented amodally as abstract symbols, with purely arbitrary associations between concepts, and are largely unrelated to the sensorimotor experience of the concepts that are expressed (Pylyshin, 1984, Shallice, 1988). An alternative view is proposed by the Embodied Semantic Theory (see, for example, Barsalou, 2008). According to the latter perspective, the meaning of a spoken or read word, corresponding to a concept, is represented in the brain by the same neural networks that underlie the perceptual and motor experiences that have been associated when experiencing its referent. In the case of words referring to actions, some studies have suggested a possible role of the mirror neuron system (Gallese et al., 1996, Gallese and Lakoff, 2005). Over the last two decades, neuroimaging and neurophysiological studies have provided a growing number of evidences in line with the so-called embodied or grounded cognition framework, with important implications for current views on concepts understanding (Fernandino and Iacoboni, 2010).

Although the involvement of the mirror neurons system in language processing remains highly speculative, at a macro-anatomical level the processing of action-related words has been shown to induce functional activation of the motor and premotor areas (Pulvermüller and Fadiga, 2010). Indeed, several neuroimaging studies have investigated the representation of action-related concepts in the human brain, showing activations in the left premotor cortex when subjects read or listened to verbs or sentences describing actions (Hauk et al., 2004, Tettamanti et al., 2005). These cerebral activations were reported to be effector-specific (e.g., mouth, hand, foot), following a roughly somatotopic distribution. Oliveri et al. (2004), using transcranial magnetic stimulation (TMS), reported an enhancement of the left primary motor cortex (M1) excitability for action verb processing compared to abstract verbs during a linguistic task. Contrasting results have been reported by Buccino et al. (2005) who showed a decreased corticospinal excitability of hand or foot motor cortical areas during passive listening of action sentences involving the hand or leg. Recently, Papeo et al. (2009) demonstrated an increased excitability of the left M1 during hand-action verbs reading, compared to non-action verbs. This increased excitability was absent during reading of action verbs related to other body parts. Analogous TMS-induced corticospinal effects of hand-action verbs reading were also found by Candidi et al. (2010).

Despite growing evidence suggesting that motor circuits are involved in language understanding, more information is needed in order to better understand the cerebral mechanisms responsible for such processing. Cortical activity related to sensorimotor control can be easily studied with electroencephalography (EEG) using the event-related desynchronization/ synchronization (ERD/ERS) method (Pfurtscheller and Lopes da Silva, 1999). Mu ERD, usually observed before and during self-paced voluntary movements (Pfurtscheller and Neuper, 1992, Leocani et al., 1997, Pfurtscheller et al., 2000), as well as during reaction time paradigms (Leocani et al., 2001), passive movements (Cassim et al., 2001, Alegre et al., 2004, Houdayer et al., 2011) or motor imagery (Pfurtscheller et al., 2006), reflects the cortical activation of motor and premotor areas involved in movement preparation and execution. A desynchronization is considered to reflect the activation of cortico-thalamo-cortical neuronal circuits (Steriade and Llinas, 1988, Pfurtscheller and Berghold, 1989, Neuper and Pfurtcheller, 1999). Mu ERD has also been widely used to study cortical activation during cognitive performances (Boiten et al., 1992, Klimesch et al., 1993, Dujardin et al., 1995, Stipacek et al., 2003). These authors demonstrated that the alpha power (particularly in the upper alpha band) decreases with increasingly demanding tasks (Stipacek et al., 2003). EEG power measures are highly reliable within individuals (Neuper et al., 2005), and alpha band appears to be the most sensitive indicator of differences among individuals, probably due to its higher inter-individuals variance as compared to other frequency bands (van Dis et al., 1979). As a consequence, mu ERD analyses may constitute a powerful tool to study the cerebral activity involved in action-related sentence processing.

In the present study, mu ERD analysis was used to address two questions. First, does motor cortex activation occur automatically when participants perform a sentence-reading task involving hand action verbs, even when the task does not require an explicit retrieval of the motor content? Second, does this cortical motor activation occur also during the reading of action verbs presented with sentential negation? Sentential negation is a universal syntactic feature of human language that reverses the truth value of any given sentence by using lexical morphemes such as “not”; e.g. “John has left” and “John has not left” (Klima, 1964, Laka, 1990, Haegeman, 1997, Zanuttini, 1997). Several recent psycholinguistic studies have investigated the influence of sentential negation on language processing, but the neurophysiological evidence is limited. Tettamanti et al. (2008) used fMRI to show that action-related sentences, compared to abstract ones, activated the left premotor and fronto-parietal areas involved in action representation. The same authors found a reduced BOLD signal activity for negative versus affirmative action-related sentences. Liuzza et al. (2011), applying paired-pulse TMS over the left M1 during passive reading of hand-action and abstract sentences, showed a difference in cortico-spinal excitability for positive and negative polarity hand-action sentences.

In the present study, EEG was recorded in young healthy volunteers during a sentence-reading task. Subjects were asked to read silently affirmative or negative short sentences, with hand-action or abstract content. When the sentence disappeared from the screen, subjects had to decide whether the verb contained in the sentence was represented in the Italian language with high, medium or low frequency (Fig. 1). We hypothesized that reading sentences referring to hand-action compared to abstract verbs would be associated to motor cortex involvement as measured by mu ERD. We also hypothesized that sentence polarity would modulate mu ERD, possibly leading to a lower amplitude of mu ERD during negative hand-action sentences reading.