Exploring the representational basis of response-effect anticipation : modality-specificity and conceptual generalization in action control

According to the ideomotor principle, action control is based on the representation of anticipated effects the actions cause (Greenwald, 1970). This notion has been investigated in many studies over the last two decades, using the response-effect compatibility (REC) paradigm, in which participants’ responses are followed by a compatible or incompatible effect (i.e., compatible effects overlap with the response on some dimension, while incompatible effects do not). Participants typically respond faster when the anticipated effect is compatible to the response compared to when it is incompatible, suggesting that the effect was anticipated and that it facilitated response selection and initiation (i.e., REC effect; e.g., Kunde, 2001). Importantly, compatibility is created based on pre-existing associations, thus, R-E overlap does not have to be learned during the experiment. For example, if a right-hand response is followed by an effect appearing on the right, the effect is compatible to the response in terms of spatial location, whereas, in contrast, a right-side effect after a left-hand response is incompatible due to the spatial mismatch between response and effect locations. Thus, this way of creating R-E compatibility is similar to the way S-R compatibility is created (e.g., Simon task). The present dissertation investigates how generic vs. specific the representation of response effects is in terms of (1) the modality of response effects in relation to response modality (i.e., response-effect modality mappings) and (2) the influence of the conceptual-semantic content of effects. According to the notion of ideomotor compatibility (Greenwald, 1972) the way a stimulus primes a response can be described by the level of similarity between S and R, and therefore vocal responses are easier to initiate after an auditory stimulus. The more generic account of modality compatibility (e.g., Stephan & Koch, 2010) explains the strong relationship of vocal-auditory modalities by pointing out that a response in vocal output modality most often leads to an auditory effect (e.g., through verbal communication), while manual responses often result in a visual outcome (i.e., typing, pointing). The theory suggests that since manual-visual and vocal-auditory input-output modality mappings often occur together in daily life, these modality mappings have a stronger connection than for example, manual-auditory and vocal-visual input-output modality mappings. Even though the core idea of modality compatibility refers to responses and their typical outcomes in terms of modalities (i.e., response-effect modality mapping), comparison of performance with compatible modality mappings (i.e., vocalauditory or manualvisual) and incompatible modality mappings (i.e., manualauditory and vocalvisual) has previously been done only in stimulus-response studies (e.g., Fintor, Stephan, & Koch, 2018; Greenwald, 1972). Studies II and III focused on the conceptual feature of the representation of anticipated response effects. Earlier studies suggest that response-effect anticipation might have a conceptual basis (e.g., Hommel, Alonso, & Fuentes, 2003; Koch & Kunde, 2002). However, evidence cannot be considered conclusive due to methodological issues (e.g., in the setting of Koch and Kunde [2002], conceptual R-E overlap was confounded with phonological overlap between response and effect in the compatible condition). Studies II and III further investigated conceptual generalization of response effects in two novel settings of the REC paradigm (Study II: bilingual REC, Study III: category-exemplar REC), in order to exclude phonological overlap in the compatible condition. Furthermore, in Study III, the conceptual link was assessed both in exogenous (stimulus-response compatibility [SRC] task) and in endogenous response priming (REC). Study I has shown that response-effect representations are modality-specific. Furthermore, that these influences are generic (i.e., they are present in single-task settings, not exclusively in multitasking settings as it had been proposed by earlier research). Studies II and III could not confirm the role of conceptually related response effects in action control: The influence of perceptual-phonological overlap was stronger. Furthermore, Keller and Koch (2006) suggested that stimulus-response compatibility and response-effect compatibility are functionally equivalent; however, since only a SRC while no reliable REC effect was found, the contrast in Study III does not confirm this assumption for conceptually related words. In summary, while the two-stage model of Elsner & Hommel (2001) claims that every effect feature has the same priming power in response-effect anticipation with no qualitative difference among the different features, the present findings call for an extension of this assumption. Furthermore, the findings on conceptual generalization suggest that some response effects might have a direct effect on the motor response (online action control), while other response effects might not be directly movement-related and might guide actions in a more indirect way (offline action control)