Sensory attenuation of action outcomes of varying amplitude and valence

https://doi.org/10.1016/j.concog.2020.103058Get rights and content

Highlights

  • Sounds of varying amplitude and valence were used to test sensory attenuation accounts.

  • Sensory amplification was observed in action outcomes of low amplitude and positive valence.

  • Sensory attenuation is a flexible effect and its direction can be reversed.

  • Explicit sense of agency was affected differently than sensory attenuation.

  • Intentional action also influenced subjective ratings of stimuli’s valence.

Abstract

Stimuli caused by self-initiated actions are perceived as less intense than those caused externally; this effect is called sensory attenuation (SA). In two experiments, we aimed to assess the impact of the amplitude of outcomes and its affective valence on SA and explicit ratings of sense of agency. This allowed us to test the predictions of the available SA frameworks and better understand the link between SA, affect, and agency. The results indicated that SA can be reversed, and such sensory amplification is driven by low-amplitude and positive-valence outcomes. We also show that intentional action influences the perceived valence of outcomes, and that modulations of explicit sense of agency are divergent from those of SA. Our study shows that valence influences the processing of the amplitude of intentional action outcomes and suggests that none of the currently available frameworks give full justice to SA’s variability.

Introduction

Self-related stimuli are processed in a specific way and elicit a different set of cognitive and neurophysiological reactions than stimuli with no self-related meaning or value (Northoff et al., 2006). An exemplary manifestation of this is perception of external stimuli. Processing of such stimuli differs between situations in which they are caused by self-initiated volitional actions and when they are caused by external forces that are passively observed by an agent. One dimension which is affected by self-action is the magnitude or amplitude of perceived outcomes. This phenomena is known as sensory attenuation (SA; also referred to as sensory suppression): self-caused events are generally perceived to be weaker as compared to physically identical events that are caused externally (Blakemore, Wolpert, & Frith, 1998). The SA effect is observable in both mild (Blakemore, Frith, & Wolpert, 1999) and nociceptive (Wang, Wang, & Luo, 2011) tactile stimulation, as well is in auditory (Weiss, Herwig, & Schütz-Bosbach, 2011) and visual modalities (Cardoso-Leite, Mamassian, Schuttz-Bosbach, & Waszak, 2010). Apart from the perceptual level, SA is also visible on the neurophysiological level: processing of self-caused events has been linked to diminished activation of the anterior cingulate cortex (Blakemore et al., 1998) and attenuated event-related potentials such as auditory N1 (Klaffehn, Baess, Kunde, & Pfister, 2019) and visual N2/P3 (Hughes & Waszak, 2011). Intriguingly, rather than attenuation, some studies also report sensory amplification of self-generated actions (e.g. Rezniket al., 2014, Simões-Franklinet al., 2011; discussed below). Due to the fact that SA is observable only in self-initiated actions, it has been has been suggested that the effect is a marker of a feeling of being the author of one’s own actions and their effects (Blakemoreet al., 1999, Gentsch and Schütz-Bosbach, 2011), or a sense of agency (SoA) (Haggard, 2017, Moore and Fletcher, 2012). SA can thus be seen as an alternative measure of SoA to the more commonly used intentional binding (Haggard et al., 2002, Moore and Obhi, 2012), although both effects seem to be based on a rather different mechanisms (Dewey and Knoblich, 2014, Hugheset al., 2013).

Even though SA is observable across senses and with different types of stimuli, its mechanism is still not well understood. Multiple factors may be involved, such as a stimulus’ psychophysical features (e.g. amplitude) and affective context. Notably, perceptual and neural SA have been shown to be driven by a distinct mechanism. In their study, Palmer, Davare, & Kilner (2016) investigated whether somatosensory evoked potential could be related to the behavioral measure of SA in a force-matching paradigm. The authors found that behavioral SA and attenuated somatosensory evoked potentials were dissociable, which suggests that a distinct neurophysiological mechanism underlies these two effects. Similarly to studies employing intentional binding (Ebert and Wegner, 2010, Majchrowicz and Wierzchoń, 2018), another line of research has shown that SA can also be dissociated from explicit judgement of SoA. This shows that implicit and explicit measures of SoA may be based on different mechanisms (Dewey and Knoblich, 2014, Gentschet al., 2012) and may also reflect distinct dimensions of SoA: implicit pre-reflective feeling and explicit declarative judgement, respectively (Synofzik, Vosgerau, & Newen, 2008).

Several cognitive frameworks have been proposed to account for the SA effect that was observed in these and other studies. However, as reviewed below, each of these frameworks leads to different predictions regarding the effect’s dependence on specific psychophysical factors. The aim of the current study is to elucidate which framework best accounts for SA’s dependence on two factors. The first is the objective amplitude (magnitude) of the outcome, which is investigated in Experiment 1. The second factor, which until now has been mostly ignored in this area of research, is the emotional valence of the outcomes, as is introduced in greater detail and directly investigated in Experiment 2.

The first account of SA, the so-called cancellation account, is based on the classic computational model of motor control (Wolpert, 1997). A key element of this account is the comparator mechanism, which compares specific signals, such as the predicted (pre-movement) bodily state, against the actual (post-movement) state. This provides the bodily feedback signal (reafference) that is used for the optimization and refinement of subsequent movements. In the context of SA, a crucial signal in this process is the efference copy of a motor command, which is used to predict and cancel out (or dampen) the sensory feedback caused by subsequent movements (Bays and Wolpert, 2007, Blakemoreet al., 2002). One specific prediction of this account is that SA is an effect of central cancellation (subtraction) applied to all self-generated stimulation that does not depend on the objective amplitude of the stimulation. That is, the cancellation is supposed to be constant regardless of the intensity of the self-generated input (Bays & Wolpert, 2007b).

More recent accounts, however, have suggested that this view is insufficient as the physical amplitude of the stimuli can interact with the SA effect. For example, an alternative SA framework, called the preactivation account, proposes that the effect results not from decreasing the incoming signal’s strength but from increasing the neural activity preceding the input (Roussel, Hughes, & Waszak, 2013). The key idea here is that self-generated stimulation is predictable, and this prediction leads to preactivation of the selected sensory areas. Such a preactivation mechanism can be linked to wider considerations on neural correlates of consciousness which have suggested that a few slow event-related potentials (such as contingent negative variation, readiness potential, or stimulus-preceding negativity) constitute a prerequisite or even proper correlate(s) of conscious processing (cf. Rutiku & Bachmann, 2017, section 2). Importantly, preactivation depends on the amplitude of the stimulation, i.e. with a higher level of activity the increase in internal response or preactivation is weaker than in the case of lower activity due to saturation (i.e. activity reaches its possible maximum). In Signal Detection Theory terms (Green & Swets, 1966), this framework predicts lower signal detection sensitivity (d′) as the stimulus-driven increase of the internal response is smaller when preactivation is present compared to when it is not, and the same is true for discrimination performance. Thus, the preactivation account suggests that the preactivation would get weaker (due to saturation) with self-generated outcomes of higher amplitudes, thus resulting in less pronounced or absent SA. Roussel and colleagues (2013) conducted an experiment using a contrast discrimination task to test the predictions of their account. They observed reduced sensitivity in congruent (i.e. predictable) as compared to incongruent (unpredictable) stimuli, but only for low-contrast stimuli. The difference between contrasts can be explained by the effect of saturation, which is more present with high-intensity stimuli. This shows that (due to differences in preactivation) SA is present at low amplitudes of stimulation, but due to saturation it disappears at higher amplitudes.

The cancellation and preactivation accounts thus propose different predictions regarding how SA depends on the outcome’s amplitude. None of these frameworks, however, account for studies reporting cases of sensory amplification (instead of attenuation) of self-generated actions. For example, in one fMRI study (Simões-Franklin et al., 2011), participants were asked to categorize the roughness of a surface by either active exploration or passive touch (when the surface was moved under their fingers). The results showed greater activation of somatosensory areas in the active vs passive condition (see also Ackerley et al., 2012). Reznik, Henkin, Schadel, and Mukamel (2014) reported similar results in an auditory modality: bilateral increase in the activity of the auditory cortices in response to self-generated sounds as compared to those generated externally. To also account for these cases of sensory amplification or reversed SA, Reznik, Henkin, Levy, and Mukamel (2015) proposed that the strength and direction of the SA effect directly depends on the predicted amplitude of the outcome. According to this (as the authors called it) expected intensity account, self-generated outcomes of low amplitudes are internally amplified, while at higher intensities the perceived strength of stimuli is attenuated. To verify this hypothesis, Reznik et al. (2015) used near-threshold and supra-threshold auditory stimuli and compared their perceived loudness in active action and passive observation task conditions. The results showed that SA occurred when active actions resulted in supra-threshold outcomes. However, this pattern was reversed in the case of near-threshold outcomes: reversed SA was visible instead. In sum, the current research provides mutually contradictive results, and the main theoretical accounts of SA lead to similarly contradictive predictions regarding the effect’s strength and direction due to the amplitude of the outcomes (see Fig. 1).

The current ambiguity in the research on SA is thus visible on the level of both empirical data and theoretical predictions. The general picture that emerges is that SA may be not as stable and static an effect as is sometimes implicitly postulated as it can vary depending on various contextual factors. While current models remain mostly agnostic as to what the most crucial factors might be, various pieces of research have provided some clues. One obvious factor is action-effect contingency and/or the predictability of outcomes (Baesset al., 2011, Doggeet al., 2019). The strength of prior beliefs of authorship have also been shown to modulate SA (Desantis, Weiss, Schütz-Bosbach, & Waszak, 2012). In a neighboring research line focusing on intentional binding as a measure of SoA, it has been shown that emotional valence and the incentive value of outcomes can strongly influence intentional binding (Di Costaet al., 2017, Takahataet al., 2012, Yoshie and Haggard, 2013). Moreover, affective content influences both stimuli-related expectancy and its subsequent processing (e.g. Barrett and Bar, 2009, Murphy and Zajonc, 1993). Thus, it is likely that affective content may modulate the SA effect but, to the best to our knowledge, until now this issue has not been studied in this context.

The main aim of the current study is to address the current ambiguity in SA accounts and research and to further investigate the dependence of SA on specific, objective characteristics of outcomes. To do so, we conducted two experiments with tasks based on previous work in this research area (Rezniket al., 2015, Rousselet al., 2013), and with varying types of stimuli and complexity of manipulations. In Experiment 1, as action effects we used pure sounds, which are typically used in psychophysical research on SA and SoA. In this way we ensured that the stimulation was similar to that used in previous studies. By varying the amplitude of the outcomes between blocks, we aimed to assess SA amplitude in these blocks, thus testing the actual dependence of the effect’s direction on objective amplitude. To quantify SA, we used direct ratings of stimulus intensity, similarly to Roussel et al. (2013), who did so in the context of visual stimuli (luminance ratings). The results of Experiment 1 indicated that standard SA is not always present despite using typical stimuli. In Experiment 2, we extended our design by manipulating an additional purportedly important factor for SA: the affective valence of outcomes. Thus, instead of simple sounds, human affective vocalizations were used across blocks. Moreover, two additional dependent measures were used: First, a subjectively perceived valence scale was employed to investigate potential action-dependent modulation on valence perception. Second, an agency rating scale was used to link explicit dimensions of SoA to the SA measure.

Section snippets

Participants

35 volunteers (8 males) participated in the study; mean age was 24.5 (SD = 4.6). The sample size was chosen to approximately match or surpass the power of a previous study that reported SA modulations due to outcome amplitude (Reznik et al., 2015). Prior to the experiment, each participant gave informed consent. Participants received 10 PLN or course credits in return for participation. All had normal or corrected-to-normal vision. The research was approved by the local research ethics

Experiment 2

It seems that the direction and strength of SA can vary between contextual conditions, yet the specific factors that modulate this effect are still unknown. Factors that have been shown to generally influence SA include the predictability of outcomes, prior causal beliefs, and freedom of choice (Borhaniet al., 2017, Desantiset al., 2012, Doggeet al., 2019). Although it has been ignored in SA research, self-relevance is closely linked to the affective content or meaning of contextual

General discussion

The current study aimed to assess whether and how the objective amplitude and affective content of an intentional action’s effect influence sensory attenuation effect’s direction and strength. In two experiments, participants either intentionally caused or passively perceived auditory outcomes which varied in amplitude; they then rated their subjectively perceived amplitude using a visual analog scale. Comparing reported amplitude in active and passive tasks allowed us to assess the SA effect,

CRediT authorship contribution statement

Bartosz Majchrowicz: Conceptualization, Methodology, Software, Investigation, Visualization, Writing - original draft, Funding acquisition. Michał Wierzchoń: Conceptualization, Methodology, Writing - review & editing, Supervision.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

The authors would like to thank Natalia Wójcik (Institute of Psychology, Jagiellonian University) for help with data acquisition. The study was supported by the Polish Ministry of Science and Higher Education under the Diamond Grant program (0097/DIA/2014/43), granted to BM.

References (65)

  • G. Northoff et al.

    Self-referential processing in our brain-A meta-analysis of imaging studies on the self

    NeuroImage

    (2006)
  • M. Pyasik et al.

    On the relation between body ownership and sense of agency: A link at the level of sensory-related signals

    Acta Psychologica

    (2018)
  • C. Roussel et al.

    A preactivation account of sensory attenuation

    Neuropsychologia

    (2013)
  • R. Rutiku et al.

    Juxtaposing the real-time unfolding of subjective experience and ERP neuromarker dynamics

    Consciousness and Cognition

    (2017)
  • K. Sandberg et al.

    Measuring consciousness: Is one measure better than the other?

    Consciousness and Cognition

    (2010)
  • A.K. Seth

    Interoceptive inference, emotion, and the embodied self

    Trends in Cognitive Sciences

    (2013)
  • M. Synofzik et al.

    Beyond the comparator model: A multifactorial two-step account of agency

    Consciousness and Cognition

    (2008)
  • C. Weiss et al.

    The self in action effects: Selective attenuation of self-generated sounds

    Cognition

    (2011)
  • C. Wilke et al.

    The valence of action outcomes modulates the perception of one’s actions

    Consciousness and Cognition

    (2012)
  • D. Wolpert

    Computational approaches to motor control

    Trends in Cognitive Sciences

    (1997)
  • M. Yoshie et al.

    Negative emotional outcomes attenuate sense of agency over voluntary actions

    Current Biology

    (2013)
  • R. Ackerley et al.

    An fMRI study on cortical responses during active self-touch and passive touch from others

    Frontiers in Behavioral Neuroscience

    (2012)
  • A. Anzulewicz et al.

    Bringing action into the picture. How action influences visual awareness

    Attention, Perception, & Psychophysics

    (2019)
  • P. Baess et al.

    Selective suppression of self-initiated sounds in an auditory stream: An ERP study

    Psychophysiology

    (2011)
  • L.F. Barrett et al.

    See it with feeling: Affective predictions during object perception

    Philosophical Transactions of the Royal Society B: Biological Sciences

    (2009)
  • D. Bates et al.

    Fitting Linear Mixed-Effects Models using lme4

    Journal of Statistical Software

    (2015)
  • P.M. Bays et al.

    Computational principles of sensorimotor control that minimize uncertainty and variability

    The Journal of Physiology

    (2007)
  • P.M. Bays et al.

    Predictive attenuation in the perception of touch

  • S.J. Blakemore et al.

    Spatio-temporal prediction modulates the perception of self-produced stimuli

    Journal of Cognitive Neuroscience

    (1999)
  • S.J. Blakemore et al.

    Central cancellation of self-produced tickle sensation

    Nature Neuroscience

    (1998)
  • Bolker, B. (2019). GLMM FAQ (retrieved from:...
  • K. Borhani et al.

    Choosing, doing, and controlling: Implicit sense of agency over somatosensory events

    Psychological Science

    (2017)
  • Cited by (5)

    • Social, affective, and non-motoric bodily cues to the Sense of Agency: A systematic review of the experience of control

      2022, Neuroscience and Biobehavioral Reviews
      Citation Excerpt :

      Tellingly, some studies found a correlation between direct and indirect measures of SoA (Caspar, Beyer et al., 2021; Caspar, De Beir et al., 2021; Schwarz et al., 2019). However, other authors reported that the same manipulations did not produce comparable modulations of explicit SoA and IB (e.g., Dewey and Knoblich, 2014; Wen et al., 2015a), and of explicit SoA and SA (Garrido-Vásquez and Rock, 2020; Majchrowicz and Wierzchoń, 2021). This may suggest that direct and indirect measures capture partially distinct processes, or that they gauge different levels of SoA.

    • Non-motor cues do not generate the perception of self-agency: A critique of cue-integration

      2022, Consciousness and Cognition
      Citation Excerpt :

      For instance, an affect state leads to intentional binding (e.g., Yoshie & Haggard, 2013) as well as it does not (e.g., Moreton, Callan, & Hughes, 2017); conceptual beliefs give rise to intentional binding (e.g., Buehner, 2012) and sometimes they will not (e.g., Antusch et al., 2020); and (non-motor) vicarious observation of some other’s action gives rise to intentional binding (e.g., Moore et al., 2013; Poonian & Cunnington, 2013) and does not give rise to it as well (e.g., Cravo, Claessens, & Baldo, 2011). Similarly, affect states sometimes lead to sensory attenuation (e.g., Gentsch et al., 2015) and some other times, they do not (e.g., Majchrowicz & Wierzchoń, 2021); and (passive or vicarious) observation of others’ action leads to sensory attenuation (e.g., Pyasik et al., 2021; Sato, 2008) and also does not lead to sensory attenuation (e.g., Kilteni et al., 2021; Thomas, Sink, & Haggard, 2013; Weiss, & Schütz-Bosbach, 2012). These contradictory subjective reports—under the context of very same voluntary agency/action states, in general, as well as in the context of very same non-motor cues, in particular—imply that the verbal reports elicited using the implicit sense of agency measures are also judgment effects rather than perceptual effects, as true sensory percepts do not change or contradict from time to time (unless there occur alterations in either the stimulus or the sensory modality).

    View full text