Elsevier

Cognition

Volume 112, Issue 3, September 2009, Pages 467-472
Cognition

Brief article
Unintended embodiment of concepts into percepts: Sensory activation boosts attention for same-modality concepts in the attentional blink paradigm

https://doi.org/10.1016/j.cognition.2009.06.003Get rights and content

Abstract

This study shows that sensory priming facilitates reports of same-modality concepts in an attentional blink paradigm. Participants had to detect and report two target words (T1 and T2) presented for 53 ms each among a series of nonwords distractors at a frequency of up to 19 items per second. SOA between target words was set to 53 ms or 213 ms, with reduced attention expected for T2 under the longer SOA (attentional blink) and for T1 under the shorter SOA (lag-1 sparing). These effects were found but reduced when the sensory modality of the concepts matched that of a sensory stimulation occurring prior to the detection trial. Hence, sensory activation increased report for same-modality concepts. This finding reveals that grounded cognition effects (1) are involved in conceptual processing as soon as a word has reached the point of lexical identification and (2) occur independent of intentional access to sensory properties of concepts.

Introduction

When presenting rapid serial visual presentation (RSVP, with up to 19 items per second), a negative effect of the first target on the second target identification is observed, an effect coined attentional blink (Raymond, Shapiro, & Arnell, 1992). Attentional blink (AB) is typically defined as interference from an initial to-be-processed target (T1) on the processing of a second target (T2) appearing within 200–500 ms after T1. The AB research has suggested the existence of a limitation of attentional capacity to process and memorize competing stimuli (identified as targets) in a rapid stream of stimuli. This effect is function of the Stimulus Onset Asynchrony (SOA). It is maximal at an SOA between targets of about 200 ms and disappears when the SOA is about 500 ms. Of importance, a reverse effect (i.e., an advantage of T2 over T1) is observed at very short SOAs (i.e., from 13 to 53 ms; Potter, Staub, & O’Connor, 2002). In that latter case, T2 often ‘blinks’ T1, reducing its report, an effect called ‘lag-1 sparing’ (e.g., Olivers & Meeter, 2008).

Of note, even when a target cannot be reported, it is often processed, identified and recognized as a meaningful stimulus (e.g., a word). The idea that meaning is extracted very quickly during word processing (i.e., as soon as a word is lexically identified) comes from studies on sentence processing using RSVP paradigms. Potter, Kroll, Yachzel, Carpenter, and Sherman (1986) showed that meaningful sentences of about 14 words can be recalled with a high accuracy even when the sequential presentation rate is about 10–12 words per second (see also Potter, Stiefbold, & Moryadas, 1998). There is also electrophysiological evidence suggesting that the meaning (measured with the N400 component) of a blinked word during the AB effect is accessed even though it may not be consciously reported (Luck, Vogel, & Shapiro, 1996).

The present study builds upon the above AB research and on the recent grounded cognition framework (e.g., Barsalou, 1999, Barsalou, 2008). The latter grounded or embodied cognition framework proposes that memory, language or judgments directly depend on simulations in sensory–motor systems that were active during the initial experience rather than on abstract symbols (i.e., amodal systems) that consist of transcriptions of neural states. We reasoned that sensory–conceptual interactions may arise as soon as a word has reached the point of lexical identification, reducing AB and lag-1 sparing effects after a sensory activation in a modality matching that of the target word.

Recent studies have started to examine the facilitation or the inhibition influence of modal perception on conceptual access. For instance, in van Dantzig et al., 2008, Vermeulen et al., 2008), participants performed alternatively a perceptual decision task and a conceptual property-verification task. Results showed better performances on the property-verification task for trials preceded by a perceptual decision in a same rather than in a different modality as that of the to-be-verified concepts. Such demonstration supports the grounded cognition view: perceptual and conceptual representations are partially based on the same systems, so that sensory priming influences conceptual access.

However, most grounded cognition effects have been reported in tasks where targets words could be clearly identified and where participants were explicitly instructed to verify or to generate sensory properties of concepts (e.g., to verify whether “Lemon” is “Yellow”). Hence, the question remains whether similar effects may be observed under much quicker exposure times (53 ms) and in the context of tasks where access to sensory properties of the concepts is unintended. Grounded cognition frameworks do not make strong a priori predictions about unintended (passive) knowledge access. Yet, empirical support can be found in neuroimaging studies of word processing. For instance, using event-related fMRI, Hauk, Johnsrude, and Pulvermüller (2004) asked participants to read passively action words pertaining to the categories of face- (e.g., to lick), arm- (e.g., to pick), and leg- (e.g., to kick). These authors showed action-specific activation in brain areas directly adjacent to, or overlapping with, areas involved during actual movement of the tongue, finger or foot.

In this research, we relied on the AB paradigm to examine our hypothesis. We crossed the modality of the sensory activation (visual vs. auditory) and the modality of the target word (word related to vision vs. audition) in the context of a RSVP of words study. We predicted a four-way interaction between SOA (short or long), Target position (T1 or T2), sensory activation (auditory or visual) and target modality (vision-related or audition-related word), with higher reports of targets words associated with AB and lag-1 sparing effects following a congruent sensory activation. In other words, we predicted higher reports of blinked targets (i.e., T2 at long SOA) for vision-related words following a visual stimulation and for audition-related words following an auditory stimulation than for vision-related words following an auditory stimulation and for audition-related words following a visual stimulation. A same sensory-matching advantage was expected for the lag-1 sparing effect, this time concerning T1 at the short SOA. No sensory-matching advantage was expected in this task for T2 at the short SOA and for T1 at the long SOA, due to ceiling effect related to conscious perception of the target.

Section snippets

Participants and design

Twenty French-speaking volunteers (10 women, mean age = 22.50 years, SD = 2.16) were paid 5€ (around 7.4$ at the time of the study) to participate. The study was introduced as concerned with attentional capacity and conformed to a 2 (sensory activation: auditory vs. visual) × 2 (Target position: T1 vs. T2) × 2 (SOA: 53 ms vs. 213 ms) × 2 (Word modality: audition-related vs. Vision-related words) fully within-subjects design.

Materials

For the visual activation, we used a visual presentation of the word “PRÊT” (READY

Results

Performance on each word in the pair was scored separately so that misspelled words and blanks were counted as errors (e.g., Davenport & Potter, 2005). Overall, 53% of the words were reported accurately which corresponds to slightly more than one word per trial (on two possible words). An analysis of variance with repeated measures (MANOVA) was used with 2 (sensorial activation: auditory vs. visual) × 2 (Target position: T1 vs. T2) x 2 (SOA: 53 ms vs. 213 ms) x 2 (Word modality: audition-related

Discussion

Consistent with Potter et al., 2002, Potter et al., 2005, an AB was observed on T2 under the long SOA whereas a lag-1 sparing effect was observed on T1 under the short SOA. However, these effects were reduced following a sensory activation in a modality matching that of the concept. These findings fit well with grounded cognition at the expense of amodal theories of concept representation. In amodal systems, abstract symbols representing knowledge have no remaining link with perception from

Conclusion

This research shows that the AB and lag-1 sparing effects are modulated by the congruency of a prior sensory activation. This important finding suggests that psychological processes linking perception to conceptual knowledge do not necessarily require intentional access to concept properties. Rather, grounded cognition may actually take place at very short exposure times even when people do not intend to represent the modal properties of concepts. Perception and conceptual knowledge may then be

References (19)

There are more references available in the full text version of this article.

Cited by (31)

  • Feeling better: Tactile verbs speed up tactile detection

    2020, Brain and Cognition
    Citation Excerpt :

    This suggests that brain sensorimotor areas may be called upon early during word lexico-semantic access (Boulenger et al., 2006, 2012; Hauk & Pulvermüller, 2004; Mollo et al., 2016; Ostarek & Huettig, 2017; Shtyrov et al., 2014) and that this activation may persist at least until 500 ms post-word onset (see, for similarly late effects, Papeo et al., 2009). The priming effect observed despite no explicit task was performed on verbs provides additional evidence that somatosensory activation occurs independently of intentional access to words’ sensory properties (Connell & Lynott, 2010; Shtyrov et al., 2014 for action words; Vermeulen et al., 2009). The results of Experiment 2 however highlight that the benefit for tactile detection may depend on the site of tactile stimulation.

  • Congruent bodily arousal promotes the constructive recognition of emotional words

    2017, Consciousness and Cognition
    Citation Excerpt :

    Previous results showed that increased bodily arousal (i.e., following a cycling session) led to improved reports of high arousal T2 words, whereas reduced bodily arousal (i.e., following a relaxation session) led to improved reports of low arousal T2 words. It can thus be suggested that arousal-congruent words enjoy facilitated access to the working memory space, increasing the possibility that they get accurately reported (Kever et al., 2015; Vermeulen, Chang, Mermillod, Pleyers, & Corneille, 2013; Vermeulen, Mermillod, Godefroid, & Corneille, 2009). In view of these promising preliminary findings, further studies are clearly needed to determine more specifically under which conditions arousal congruency effects can be observed.

  • Body-part specific interactions of action verb processing with motor behaviour

    2017, Behavioural Brain Research
    Citation Excerpt :

    Since the motor system is not only well characterised neuroanatomically and neurophysiologically, but also generates specific output, i.e. motor behaviour, this offers intriguing research questions with respect to the functional interaction of motor execution and language processing. Research on other features in language processing, such as different sensory properties, has also provided compelling evidence for the characteristics of semantic processing and embodied cognition [8–10]. Nevertheless, motor language is special due to its link to motor execution, which can be readily studied in behavioural experiments.

  • Is the difference between right and left ATLs due to the distinction between general and social cognition or between verbal and non-verbal representations?

    2015, Neuroscience and Biobehavioral Reviews
    Citation Excerpt :

    Furthermore, activating sensory-motor representations by pictures of manipulable objects (Helbig et al., 2006), action movies (Helbig et al., 2010) or actual hand or leg movements (Shebani and Pulvermüller, 2013) influences object naming or verbal working memory. Finally, sensory stimulation incurs performance on corresponding features in conceptual tasks (Vermeulen et al., 2008, 2009). In any case, I do not think that ATL function is well understood, if possible interactions with the sensory-motor systems are disregarded (e.g. Sim et al., 2014) and I have already acknowledged in Section 1, that hybrid models of conceptual knowledge, which propose an interaction of ATL with the sensory-motor systems are very consistent with results of the present review.

View all citing articles on Scopus
View full text