This article was downloaded by: [University of Birmingham] On: 09 January 2015, At: 16:57 Publisher: Routledge Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK
Cognitive Neuroscience Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/pcns20
Modulation of somatosensory perception by motor intention a
b
a
a
b
Amy Parkinson , Sarah Plukaard , Sally L. Pears , Roger Newport , Chris Dijkerman & Stephen R. Jackson
a c
a
University of Nottingham , Nottingham, UK
b
University of Utrecht , Utrecht, The Netherlands
c
University of Nottingham, Nottingham, UK, and Korea University , Seoul, South Korea Published online: 25 Jan 2011.
To cite this article: Amy Parkinson , Sarah Plukaard , Sally L. Pears , Roger Newport , Chris Dijkerman & Stephen R. Jackson (2011) Modulation of somatosensory perception by motor intention, Cognitive Neuroscience, 2:1, 47-56, DOI: 10.1080/17588928.2010.525627 To link to this article: http://dx.doi.org/10.1080/17588928.2010.525627
PLEASE SCROLL DOWN FOR ARTICLE Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content. This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http:// www.tandfonline.com/page/terms-and-conditions
COGNITIVE NEUROSCIENCE, 2011, 2 (1), 47–56
Modulation of somatosensory perception by motor intention
PCNS
Amy Parkinson1, Sarah Plukaard2, Sally L. Pears1, Roger Newport1, Chris Dijkerman2, and Stephen R. Jackson1,3
Modulation of somatosensory perception
1
University of Nottingham, Nottingham, UK University of Utrecht, Utrecht, The Netherlands 3 University of Nottingham, Nottingham, UK, and Korea University, Seoul, South Korea
Downloaded by [University of Birmingham] at 16:57 09 January 2015
2
The intention to execute a movement can modulate our perception of sensory events; however, theoretical accounts of these effects, and also empirical data, are often contradictory. We investigated how perception of a somatosensory stimulus differed according to whether it was delivered to a limb being prepared for movement or to a nonmoving limb. Our results demonstrate that individuals perceive a somatosensory stimulus delivered to the “moving” limb as occurring significantly later than when an identical stimulus is delivered to a “nonmoving” limb. Furthermore, human brain imaging (fMRI) analyses demonstrate that this modulation is accompanied by a significant decrease in BOLD signal in the right parietal operculum (SII) for stimuli delivered to the moving limb. These results indicate that during movement preparation a network of premotor brain areas may facilitate movement execution by attenuating the processing of behaviorally irrelevant signals within higher-order secondary somatosensory (SII) areas.
Keywords: fMRI; Functional connectivity; Motor prediction; Somatosensory function; Motor intention.
INTRODUCTION The intention to execute a movement can modulate our perception of sensory events (Blakemore, Wolpert, & Frith, 1998; Burton et al. 1999; Chapman, Bushnell, Miron, Duncan, & Lund, 1987; Duhamel, Colby, & Goldberg, 1992; Nelson, 1987; Rizzolatti, Riggio, Dascola, & Umilta, 1987; Rorden, Greene, Sasine, & Baylis, 2002; Ross, Morrone, & Burr, 1997; Yarrow, Haggards, Heal, Brown, & Rothwell, 2001). The intention to execute a saccadic eye movement enhances perceptual processing for visual (Deubel & Schneider, 1996; Kowler, Anderson, Dosher, & Blaser, 1995) and somatosensory (Rorden et al., 2002) stimuli appearing in the vicinity of a saccade target.
Such findings have been interpreted as support for premotor accounts of attention in which stimuli at attended locations are processed more rapidly (Rorden et al., 2002), or are perceived to have occurred earlier (“prior entry”), than stimuli appearing at unattended locations (Yates & Nicholls, 2009). Similarly theoretical accounts of tactile object recognition have emphasized the perceptual benefits of “active” touch during exploratory hand movements, and human brain imaging studies have confirmed that the response to somatosensory stimulation observed in primary (SI) and secondary (SII) somatosensory cortex is greater during active compared to passive finger movements (Huttunen et al., 1996; Mima et al., 1999).
Correspondence should be addressed to: Stephen Jackson, School of Psychology, The University of Nottingham, University Park, Nottingham NG7 2RD, UK. E-mail: [email protected] Supplementary material published alongside this article at www. psypress.com/cognitiveneuroscience This work was supported in part by the National Research Foundation of Korea’s WCU (World Class University) program funded by the Ministry of Education, Science and Technology (R31-2008-000-10008-0). We are grateful to Antonia Hamilton for helpful comments.
© 2010 Psychology Press, an imprint of the Taylor & Francis Group, an Informa business www.psypress.com/cognitiveneuroscience DOI: 10.1080/17588928.2010.525627
Downloaded by [University of Birmingham] at 16:57 09 January 2015
48
PARKINSON ET AL.
By contrast, forward model (FM) accounts of motor control have postulated that whenever a motor command is issued, an efference copy of that command is passed to the appropriate FM predictor, which then generates an estimate of the sensory consequences for that movement (Von Holst & Mittelstaedt, 1954; Wolpert, 1997; Wolpert & Ghahramani, 2000). One use for such estimates may be to filter sensory information, allowing the organism to cancel or attenuate sensory signals that arise as a result of selfgenerated movements (reafference), and thereby focus processing resources on externally generated sensations (exafference) that cannot be predicted from the efference copy signal (Blakemore et al., 1998; Frith, Blakemore, & Wolpert, 2000; Von Holst & Mittelstaedt, 1954; Wolpert, 1997; Wolpert & Ghahramani, 2000). Consistent with this proposal there is now substantial evidence for the attenuation of somatosensory events that accompany self-generated movements (e.g., Blakemore et al., 1998; Chapman et al., 1987; Green, 2009; Nelson, 1987). While the attentional and FM accounts appear inconsistent, it is important to note that the central sensory gating proposed by the FM account may in fact be entirely consistent with attentional selection, if the sensory consequences of self-generated movement are largely irrelevant to behavior, and FM estimates lead primarily to the attenuation of task-irrelevant stimuli. It has been suggested that during movement it may be beneficial to attenuate somatosensory stimuli in order to better monitor proprioceptive signals (Nelson, 1996).
as a fixation and “Go” stimulus. A loudspeaker positioned on the table surface directly beneath the LED delivered white noise throughout each trial. Somatosensory stimulation to each arm was delivered by a solenoid tapper positioned 5 cm from the ulnar styloid process. Design This study involved two independent variables: the stimulated arm and the SMI. Somatosensory stimulation could be delivered to the moving arm or to the nonmoving arm. Note that the moving arm (right or left) was counterbalanced across participants. SMI consisted of the time in milliseconds between the onset of the tactile stimulus and the onset of the arm movement. The dependent measure was the participants’ verbal report of their perception of the temporal sequence of the stimulus and movement (TOJ). Procedure On each trial participants were cued to execute a unimanual reaching movement from a fixed starting point to a fixed target location, while maintaining fixation throughout (half the participants used their right hand and half used their left hand). At a predetermined time (0–150 ms before or after predicted movement onset), a 5 ms “tap” was delivered to either the moving or the nonmoving forearm. The participant’s task was, having completed the reaching movement, to verbally report whether the “tap” was delivered before or after they initiated the movement.
METHODS Analysis The methods used in this study are summarized here (additional details are given in the Supplementary material available online).
Behavioral (TOJ) task Subjects Eight naïve participants (three males, age range 23–29 years) gave informed consent and participated in this study. Seven of the participants were righthanded and all had normal or fully corrected vision.
Trials with response times greater than 1000 ms or shorter than 100 ms were excluded from the analyses. The proportion of “after” responses reported for each SMI bin was calculated for each participant separately for stimuli delivered to their moving and nonmoving arms. A logistic function was fitted to these data and PSS and JND estimates calculated.
FMRI study Participants
Apparatus Reaching movement onsets were measured using a microswitch; a red/green LED (1 cm diameter) served
Sixteen right-handed adults (six male, age range 20–33 years, mean age 25 years) participated in this study.
MODULATION OF SOMATOSENSORY PERCEPTION
Downloaded by [University of Birmingham] at 16:57 09 January 2015
Procedure Somatosensory stimulation was delivered to each forearm by pneumatically controlled, spring-loaded steel rods. Each tap was of 50 ms duration. In a stimulation alone (no movement) run, stimulation was randomly alternated between left and right arms. This run was used to define functional VOIs that are activated in the absence of movement preparation or execution. In a separate series of three MRI runs, participants were cued to make reaching movements between two buttons with their right index finger in response to a visual stimulus. At a point close to movement onset (−200 to +200 ms), the stimulation was presented randomly to one of the arms. Participants performed a total of 180 trials. MR imaging Structural and functional MRI data were acquired on a 3 T Philips Achieva Scanner using an 8-channel SENSE head coil. High resolution T1-weighted structural images were acquired using a magnetization prepared gradient echo sequence (MPRAGE, 256 slices, FOV = 256 mm, 160 transversal slices) with a resolution of 1 mm isotropic. Functional scans acquired T2*-weighted BOLD images with a voxel size of 2 × 2 × 3 mm, 30 slices, TR = 2400 ms and TE = 40 ms. Slices were contiguous and taken in a descending order. Further details of MR image preprocessing and data analysis are provided in the online supplementary material.
RESULTS AND DISCUSSION We investigated how perception of a somatosensory stimulus differed when it was delivered to a nonmoving limb or a limb that was being prepared for movement. We had individuals perform a temporal-order-judgment (TOJ) task in which they executed a reaching movement using one or other arm (counterbalanced across participants), and received brief (5 ms) punctate, above-threshold, somatosensory stimulation to either their moving or nonmoving limb. The time of somatosensory stimulation was determined by each individual’s average response time (RT) history (see Methods), and stimulation could occur either before movement onset or during the movement. After each trial, the participant judged whether the somatosensory stimulus occurred before or after movement onset. The mean proportion of “after” responses in this TOJ task was calculated for each participant, and for each stimulated arm condition, over a range of SMIs. Logistic
49
functions were then fitted separately to these data, and point-of-subjective-simultaneity (PSS) and justnoticeable-difference (JND) estimates calculated. Note that the hypothesized effects of movement intention on the PSS differ according to whether movement intention benefits somatosensory processing on the limb that is being prepared for movement, as predicted by attentional “prior entry” accounts, or leads to suppression of somatosensory processing on the to-be-moved limb, as predicted by FM accounts of motor control. These hypotheses are illustrated in Figure 1. Preliminary analyses confirmed that there was no effect of the arm (i.e. right vs. left arm) used for reaching, F(1, 6) = 0.1, p > .1, and there was no interaction between reaching arm and stimulated arm, F(1, 6) = 0.7, p > .1. Data from four representative subjects are presented in Figure 2A, and for the entire group in Figure 2B. These data indicate that the PSS for stimuli delivered to the moving limb was more negative than that for stimuli delivered to the nonmoving limb (means: moving limb = −76.4 (±64.4) ms; nonmoving limb = −11.1 (±32.7) ms; F(1, 6) = 8.3, p < .05). This finding runs counter to attentional or prior-entry accounts that predict speeded processing of stimuli delivered to the moving limb (Yates & Nicholls, 2009) and thus a positive (rightward) shift of the PSS (Figure 1A). However, it is consistent with FM accounts of motor control that predict attenuated processing on the moving limb (Blakemore et al., 1998; Chapman et al., 1987; Nelson, 1987) and thus more negative PSS values for that limb (Figure 1B). Planned comparisons revealed that whereas the PSS for the nonmoving limb did not differ from 0 ms, t(7) = −0.9, p > .1, the PSS for the moving limb was significantly different from 0 ms, t(7) = 3.4, p = .01. Analyses of the JND data revealed that there were no significant differences in JND between the moving and nonmoving limbs (means: moving arm = 62 ms, nonmoving arm = 84 ms; p > .1). One explanation of these data is that the perceived time of the movement onset is altered when the tactile stimulus is delivered to the about-to-be-moved arm relative to the nonmoving arm. However, unpublished studies that we have conducted, using a two-point discrimination threshold (2PDT) detection paradigm to quantify changes in 2PDT ahead of movement onset, suggest that this account is unlikely. Specifically, these studies indicate that 2PDT sensitivity for stimuli delivered to the about-to-be-moved arm decreases shortly before to movement onset, but this is not the case when identical stimuli are delivered to the nonmoving arm (S. L. Pears, unpublished PhD thesis).
50
PARKINSON ET AL.
(A)
(B) 1.0
1.0
Moving arm
Non-moving arm Proportion 0.5 ‘after’ responses
Proportion ‘after’ 0.5 responses
Moving arm
Downloaded by [University of Birmingham] at 16:57 09 January 2015
0.0
Non-moving arm 0.0
–200
0
200
Stimulus–Movement interval
–200
0
200
Stimulus–Movement interval
Figure 1. (A) Illustrates the predicted effect on the PSS if attending to the limb being prepared for movement leads to speeded processing of somatosensory events delivered to that limb compared to the nonmoving limb (Wolpert & Ghahramani, 2000). In this case the PSS for the moving arm is shifted rightward of the PSS for the nonmoving arm, indicating that events that occur at the same time, relative to movement onset, on the moving and nonmoving arms are nevertheless judged by the participant to have occurred earlier on the moving arm. (B) This pattern would be reversed if movement preparation led to suppression of somatosensory processing on the limb being prepared for movement (Chapman et al., 1987). In this case the PSS for the moving arm would shift leftwards of the PSS for the nonmoving arm, indicating that a stimulus must be delivered earlier to the moving arm than to the nonmoving arm for these two events to be perceived as occurring simultaneously.
An alternative, and to our mind more convincing, explanation of our data is that somatosensory stimuli delivered to the moving limb are perceived to have occurred later than when identical stimuli are delivered to the nonmoving limb. This account is consistent with the proposal that somatosensation is attenuated on the about-to-be-moved limb immediately before and during movement. To investigate the neural basis for this effect we repeated this experimental task while functional brain activity was measured using functional magnetic resonance imaging (fMRI). The blood-oxygen-level demand (BOLD) signal measured in fMRI studies does not provide an unambiguous measure of brain function; however, we hypothesized that central suppression or attenuation of somatosensation on the moving limb would likely lead to a relative reduction in BOLD signal in brain areas associated with somatosensation. Behavioral data collected in the MRI scanner were analyzed as before. The proportion of “after” responses was calculated for each participant at each SMI for each limb and a logistic function was fitted to these data. Note that all somatosensory stimuli were substantially above threshold and easily detected by all individuals. The mean PSS values from the individually fitted logistic functions for the moving and nonmoving limbs were: moving arm = −53 (± 103)
ms; nonmoving arm = 2 (± 64) ms. This difference was tested using a nonparametric permutation test which confirmed that the difference in median values of −32.6 ms was statistically significant (p < .05) (see Supplementary material for details). These data again confirm that, consistent with FM accounts, participants judged that somatosensory stimuli delivered to their moving limb occurred later than when identical stimuli were delivered to the nonmoving limb. Brain imaging data were analyzed using standard techniques within BrainVoyagerQX software (see Methods). Four predictors were defined as follows: (1) stimulation of the right arm in the absence of movement; (2) stimulation of the left arm in the absence of movement, (3) stimulation to the right (moving) arm; and (4) stimulation to the left (nonmoving) arm with movement of the right arm. A random effects general linear model (GLM) analysis was carried out to analyze effects at the group level, and the contrast of somatosensory stimulation alone (Predictors 1 and 2) vs. no stimulation was used to determine four functionally defined volumes of interest (VOIs) that were activated by somatosensory stimulation in the absence of overt movements. The analyses revealed increased BOLD activation bilaterally within the parietal operculum (which includes SII; Eickhoff, Amunts, Mohlberg, & Zilles, 2006) and bilaterally
MODULATION OF SOMATOSENSORY PERCEPTION
S1
0.5
0
prop. "after" res pons es
–175 1
0
0.8 0.6 0.4 0.2 0
0
1
S2
0.5
0
175
S3
–175
–175 1
0
175
0
175
S4
0.8 0.6 0.4 0.2 0
175
–175
(B) 1
0.8 prop. "after" responses
Downloaded by [University of Birmingham] at 16:57 09 January 2015
prop. "after" res pons es
1
prop. "after" res pons es
prop. "after" res pons es
(A)
0.6
0.4
0.2
0 –175
0
175
Figure 2. (A) Illustrates psychometric functions fitted to the TOJs of four representative participants. Data from the two movement conditions are displayed as follows: = moving arm; = nonmoving arm. Brief somatosensory stimulation was delivered to either the limb being readied for movement or the alternate limb (counterbalanced across individuals). On each trial the participants reported a TOJ between the onset of their movement and the onset of the somatosensory stimulation. Psychometric functions are fitted separately for the moving and nonmoving arm. For each individual the psychometric function, and in particular the PSS (see Methods), for the moving limb is shifted leftwards, consistent with perception of the somatosensory stimulus being delayed for the moving limb. (B) Illustrates TOJ data for the entire group. Data from the two movement conditions are displayed as follows: = moving arm; = nonmoving arm. This confirms that the pattern seen in individual participants is replicated in the group data. Where the PSS for the moving arm is shifted leftwards and is more negative than the PSS for the nonmoving arm, it is the case that a stimulus must be delivered earlier to the moving arm than to the nonmoving arm for these two events to be perceived as occurring simultaneously.
51
within the insular cortex (Table 1, Figure 3) (threshold Z = 3.29, puncorrected < .001). These findings are consistent with previous brain imaging studies of somatosensation (Burton, 2001; Olausson et al., 2002). It should be noted that the arm region of SI was also activated bilaterally as expected, but only if a less conservative statistical threshold (puncorrected < .05) was adopted. Analyses of individual data suggest that this may be due to variability in the location of SI activations following stimulation of the forearm. For this reason SI responses will not be considered further. Previous studies have shown that ipsilateral SII responses depend on projections originating in contralateral SI and SII (Burton, 2001). Additional analyses confirmed that unilateral somatosensory stimulation produced bilateral BOLD activation in both the parietal operculum and insula VOIs (see Supplementary material for details). The size of the activation clusters observed in the parietal operculum make it rather difficult to precisely localize these activations to any specific subdivision of the parietal operculum (e.g., OP1, OP4); however, it is likely that our activations coincide with area PV, which has been linked to sensorimotor integration (Burton, 2001). The influence of movement preparation on somatosensory responses can be investigated most directly by contrasting, for each of the VOIs identified in the stimulation alone (no movement) condition, how movement preparation modulates BOLD response in that VOI for stimulation to delivered to the limb being prepared for movement compared to when identical stimulation is delivered to the nonmoving limb. To address this issue we ran a three-way ANOVA with the following factors: VOI (OP vs. INS); Limb (Moving vs. Nonmoving); and Hemisphere (Left vs. Right) (note that only trials with a negative SMI, i.e., stimulation was delivered before movement onset, were entered into these analyses). The ANOVA revealed that for the OP VOI there was a significant Limb by Hemisphere interaction effect, F(1, 11) = 9.7, p < .01, that was absent for the INS VOI, F(1, 11) = 0.25, p = .6). The basis of this interaction effect was that for the VOI in the right hemisphere OP, i.e. ipsilateral to the moving limb, the BOLD response to a stimulus delivered to the moving limb was significantly reduced compared to the BOLD response to a stimulus on the nonmoving limb (Figure 4A) (means: moving limb = 7.1; nonmoving limb = 9.9; t(11) = −3.1, p < .01) (Figure 4B, 4C). This effect was not apparent for the left hemisphere OP VOI, t(11) = 1.2, p > .1. Previous studies have demonstrated that the parietal operculum (OP) and insular (INS) cortex invariably
52
PARKINSON ET AL. TABLE 1 Significant BOLD activations present during somatosensory stimulation in the absence of movement
Contrast Stimulation alone
Region
Talairach (peak activation)
Z-value
Cluster size (voxels)
Left parietal operculum (SII) region Right parietal operculum (SII) region Left insular cortex Right insular cortex
−57–31 13 54 −22 16 −36 −19 10 48 −7 4
6.7 6.04 5.77 6.31
3034 1490 2022 623
Downloaded by [University of Birmingham] at 16:57 09 January 2015
Notes: Activations were initially thresholded at a value corresponding to q (FDR)
Cognitive Neuroscience Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/pcns20
Modulation of somatosensory perception by motor intention a
b
a
a
b
Amy Parkinson , Sarah Plukaard , Sally L. Pears , Roger Newport , Chris Dijkerman & Stephen R. Jackson
a c
a
University of Nottingham , Nottingham, UK
b
University of Utrecht , Utrecht, The Netherlands
c
University of Nottingham, Nottingham, UK, and Korea University , Seoul, South Korea Published online: 25 Jan 2011.
To cite this article: Amy Parkinson , Sarah Plukaard , Sally L. Pears , Roger Newport , Chris Dijkerman & Stephen R. Jackson (2011) Modulation of somatosensory perception by motor intention, Cognitive Neuroscience, 2:1, 47-56, DOI: 10.1080/17588928.2010.525627 To link to this article: http://dx.doi.org/10.1080/17588928.2010.525627
PLEASE SCROLL DOWN FOR ARTICLE Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content. This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http:// www.tandfonline.com/page/terms-and-conditions
COGNITIVE NEUROSCIENCE, 2011, 2 (1), 47–56
Modulation of somatosensory perception by motor intention
PCNS
Amy Parkinson1, Sarah Plukaard2, Sally L. Pears1, Roger Newport1, Chris Dijkerman2, and Stephen R. Jackson1,3
Modulation of somatosensory perception
1
University of Nottingham, Nottingham, UK University of Utrecht, Utrecht, The Netherlands 3 University of Nottingham, Nottingham, UK, and Korea University, Seoul, South Korea
Downloaded by [University of Birmingham] at 16:57 09 January 2015
2
The intention to execute a movement can modulate our perception of sensory events; however, theoretical accounts of these effects, and also empirical data, are often contradictory. We investigated how perception of a somatosensory stimulus differed according to whether it was delivered to a limb being prepared for movement or to a nonmoving limb. Our results demonstrate that individuals perceive a somatosensory stimulus delivered to the “moving” limb as occurring significantly later than when an identical stimulus is delivered to a “nonmoving” limb. Furthermore, human brain imaging (fMRI) analyses demonstrate that this modulation is accompanied by a significant decrease in BOLD signal in the right parietal operculum (SII) for stimuli delivered to the moving limb. These results indicate that during movement preparation a network of premotor brain areas may facilitate movement execution by attenuating the processing of behaviorally irrelevant signals within higher-order secondary somatosensory (SII) areas.
Keywords: fMRI; Functional connectivity; Motor prediction; Somatosensory function; Motor intention.
INTRODUCTION The intention to execute a movement can modulate our perception of sensory events (Blakemore, Wolpert, & Frith, 1998; Burton et al. 1999; Chapman, Bushnell, Miron, Duncan, & Lund, 1987; Duhamel, Colby, & Goldberg, 1992; Nelson, 1987; Rizzolatti, Riggio, Dascola, & Umilta, 1987; Rorden, Greene, Sasine, & Baylis, 2002; Ross, Morrone, & Burr, 1997; Yarrow, Haggards, Heal, Brown, & Rothwell, 2001). The intention to execute a saccadic eye movement enhances perceptual processing for visual (Deubel & Schneider, 1996; Kowler, Anderson, Dosher, & Blaser, 1995) and somatosensory (Rorden et al., 2002) stimuli appearing in the vicinity of a saccade target.
Such findings have been interpreted as support for premotor accounts of attention in which stimuli at attended locations are processed more rapidly (Rorden et al., 2002), or are perceived to have occurred earlier (“prior entry”), than stimuli appearing at unattended locations (Yates & Nicholls, 2009). Similarly theoretical accounts of tactile object recognition have emphasized the perceptual benefits of “active” touch during exploratory hand movements, and human brain imaging studies have confirmed that the response to somatosensory stimulation observed in primary (SI) and secondary (SII) somatosensory cortex is greater during active compared to passive finger movements (Huttunen et al., 1996; Mima et al., 1999).
Correspondence should be addressed to: Stephen Jackson, School of Psychology, The University of Nottingham, University Park, Nottingham NG7 2RD, UK. E-mail: [email protected] Supplementary material published alongside this article at www. psypress.com/cognitiveneuroscience This work was supported in part by the National Research Foundation of Korea’s WCU (World Class University) program funded by the Ministry of Education, Science and Technology (R31-2008-000-10008-0). We are grateful to Antonia Hamilton for helpful comments.
© 2010 Psychology Press, an imprint of the Taylor & Francis Group, an Informa business www.psypress.com/cognitiveneuroscience DOI: 10.1080/17588928.2010.525627
Downloaded by [University of Birmingham] at 16:57 09 January 2015
48
PARKINSON ET AL.
By contrast, forward model (FM) accounts of motor control have postulated that whenever a motor command is issued, an efference copy of that command is passed to the appropriate FM predictor, which then generates an estimate of the sensory consequences for that movement (Von Holst & Mittelstaedt, 1954; Wolpert, 1997; Wolpert & Ghahramani, 2000). One use for such estimates may be to filter sensory information, allowing the organism to cancel or attenuate sensory signals that arise as a result of selfgenerated movements (reafference), and thereby focus processing resources on externally generated sensations (exafference) that cannot be predicted from the efference copy signal (Blakemore et al., 1998; Frith, Blakemore, & Wolpert, 2000; Von Holst & Mittelstaedt, 1954; Wolpert, 1997; Wolpert & Ghahramani, 2000). Consistent with this proposal there is now substantial evidence for the attenuation of somatosensory events that accompany self-generated movements (e.g., Blakemore et al., 1998; Chapman et al., 1987; Green, 2009; Nelson, 1987). While the attentional and FM accounts appear inconsistent, it is important to note that the central sensory gating proposed by the FM account may in fact be entirely consistent with attentional selection, if the sensory consequences of self-generated movement are largely irrelevant to behavior, and FM estimates lead primarily to the attenuation of task-irrelevant stimuli. It has been suggested that during movement it may be beneficial to attenuate somatosensory stimuli in order to better monitor proprioceptive signals (Nelson, 1996).
as a fixation and “Go” stimulus. A loudspeaker positioned on the table surface directly beneath the LED delivered white noise throughout each trial. Somatosensory stimulation to each arm was delivered by a solenoid tapper positioned 5 cm from the ulnar styloid process. Design This study involved two independent variables: the stimulated arm and the SMI. Somatosensory stimulation could be delivered to the moving arm or to the nonmoving arm. Note that the moving arm (right or left) was counterbalanced across participants. SMI consisted of the time in milliseconds between the onset of the tactile stimulus and the onset of the arm movement. The dependent measure was the participants’ verbal report of their perception of the temporal sequence of the stimulus and movement (TOJ). Procedure On each trial participants were cued to execute a unimanual reaching movement from a fixed starting point to a fixed target location, while maintaining fixation throughout (half the participants used their right hand and half used their left hand). At a predetermined time (0–150 ms before or after predicted movement onset), a 5 ms “tap” was delivered to either the moving or the nonmoving forearm. The participant’s task was, having completed the reaching movement, to verbally report whether the “tap” was delivered before or after they initiated the movement.
METHODS Analysis The methods used in this study are summarized here (additional details are given in the Supplementary material available online).
Behavioral (TOJ) task Subjects Eight naïve participants (three males, age range 23–29 years) gave informed consent and participated in this study. Seven of the participants were righthanded and all had normal or fully corrected vision.
Trials with response times greater than 1000 ms or shorter than 100 ms were excluded from the analyses. The proportion of “after” responses reported for each SMI bin was calculated for each participant separately for stimuli delivered to their moving and nonmoving arms. A logistic function was fitted to these data and PSS and JND estimates calculated.
FMRI study Participants
Apparatus Reaching movement onsets were measured using a microswitch; a red/green LED (1 cm diameter) served
Sixteen right-handed adults (six male, age range 20–33 years, mean age 25 years) participated in this study.
MODULATION OF SOMATOSENSORY PERCEPTION
Downloaded by [University of Birmingham] at 16:57 09 January 2015
Procedure Somatosensory stimulation was delivered to each forearm by pneumatically controlled, spring-loaded steel rods. Each tap was of 50 ms duration. In a stimulation alone (no movement) run, stimulation was randomly alternated between left and right arms. This run was used to define functional VOIs that are activated in the absence of movement preparation or execution. In a separate series of three MRI runs, participants were cued to make reaching movements between two buttons with their right index finger in response to a visual stimulus. At a point close to movement onset (−200 to +200 ms), the stimulation was presented randomly to one of the arms. Participants performed a total of 180 trials. MR imaging Structural and functional MRI data were acquired on a 3 T Philips Achieva Scanner using an 8-channel SENSE head coil. High resolution T1-weighted structural images were acquired using a magnetization prepared gradient echo sequence (MPRAGE, 256 slices, FOV = 256 mm, 160 transversal slices) with a resolution of 1 mm isotropic. Functional scans acquired T2*-weighted BOLD images with a voxel size of 2 × 2 × 3 mm, 30 slices, TR = 2400 ms and TE = 40 ms. Slices were contiguous and taken in a descending order. Further details of MR image preprocessing and data analysis are provided in the online supplementary material.
RESULTS AND DISCUSSION We investigated how perception of a somatosensory stimulus differed when it was delivered to a nonmoving limb or a limb that was being prepared for movement. We had individuals perform a temporal-order-judgment (TOJ) task in which they executed a reaching movement using one or other arm (counterbalanced across participants), and received brief (5 ms) punctate, above-threshold, somatosensory stimulation to either their moving or nonmoving limb. The time of somatosensory stimulation was determined by each individual’s average response time (RT) history (see Methods), and stimulation could occur either before movement onset or during the movement. After each trial, the participant judged whether the somatosensory stimulus occurred before or after movement onset. The mean proportion of “after” responses in this TOJ task was calculated for each participant, and for each stimulated arm condition, over a range of SMIs. Logistic
49
functions were then fitted separately to these data, and point-of-subjective-simultaneity (PSS) and justnoticeable-difference (JND) estimates calculated. Note that the hypothesized effects of movement intention on the PSS differ according to whether movement intention benefits somatosensory processing on the limb that is being prepared for movement, as predicted by attentional “prior entry” accounts, or leads to suppression of somatosensory processing on the to-be-moved limb, as predicted by FM accounts of motor control. These hypotheses are illustrated in Figure 1. Preliminary analyses confirmed that there was no effect of the arm (i.e. right vs. left arm) used for reaching, F(1, 6) = 0.1, p > .1, and there was no interaction between reaching arm and stimulated arm, F(1, 6) = 0.7, p > .1. Data from four representative subjects are presented in Figure 2A, and for the entire group in Figure 2B. These data indicate that the PSS for stimuli delivered to the moving limb was more negative than that for stimuli delivered to the nonmoving limb (means: moving limb = −76.4 (±64.4) ms; nonmoving limb = −11.1 (±32.7) ms; F(1, 6) = 8.3, p < .05). This finding runs counter to attentional or prior-entry accounts that predict speeded processing of stimuli delivered to the moving limb (Yates & Nicholls, 2009) and thus a positive (rightward) shift of the PSS (Figure 1A). However, it is consistent with FM accounts of motor control that predict attenuated processing on the moving limb (Blakemore et al., 1998; Chapman et al., 1987; Nelson, 1987) and thus more negative PSS values for that limb (Figure 1B). Planned comparisons revealed that whereas the PSS for the nonmoving limb did not differ from 0 ms, t(7) = −0.9, p > .1, the PSS for the moving limb was significantly different from 0 ms, t(7) = 3.4, p = .01. Analyses of the JND data revealed that there were no significant differences in JND between the moving and nonmoving limbs (means: moving arm = 62 ms, nonmoving arm = 84 ms; p > .1). One explanation of these data is that the perceived time of the movement onset is altered when the tactile stimulus is delivered to the about-to-be-moved arm relative to the nonmoving arm. However, unpublished studies that we have conducted, using a two-point discrimination threshold (2PDT) detection paradigm to quantify changes in 2PDT ahead of movement onset, suggest that this account is unlikely. Specifically, these studies indicate that 2PDT sensitivity for stimuli delivered to the about-to-be-moved arm decreases shortly before to movement onset, but this is not the case when identical stimuli are delivered to the nonmoving arm (S. L. Pears, unpublished PhD thesis).
50
PARKINSON ET AL.
(A)
(B) 1.0
1.0
Moving arm
Non-moving arm Proportion 0.5 ‘after’ responses
Proportion ‘after’ 0.5 responses
Moving arm
Downloaded by [University of Birmingham] at 16:57 09 January 2015
0.0
Non-moving arm 0.0
–200
0
200
Stimulus–Movement interval
–200
0
200
Stimulus–Movement interval
Figure 1. (A) Illustrates the predicted effect on the PSS if attending to the limb being prepared for movement leads to speeded processing of somatosensory events delivered to that limb compared to the nonmoving limb (Wolpert & Ghahramani, 2000). In this case the PSS for the moving arm is shifted rightward of the PSS for the nonmoving arm, indicating that events that occur at the same time, relative to movement onset, on the moving and nonmoving arms are nevertheless judged by the participant to have occurred earlier on the moving arm. (B) This pattern would be reversed if movement preparation led to suppression of somatosensory processing on the limb being prepared for movement (Chapman et al., 1987). In this case the PSS for the moving arm would shift leftwards of the PSS for the nonmoving arm, indicating that a stimulus must be delivered earlier to the moving arm than to the nonmoving arm for these two events to be perceived as occurring simultaneously.
An alternative, and to our mind more convincing, explanation of our data is that somatosensory stimuli delivered to the moving limb are perceived to have occurred later than when identical stimuli are delivered to the nonmoving limb. This account is consistent with the proposal that somatosensation is attenuated on the about-to-be-moved limb immediately before and during movement. To investigate the neural basis for this effect we repeated this experimental task while functional brain activity was measured using functional magnetic resonance imaging (fMRI). The blood-oxygen-level demand (BOLD) signal measured in fMRI studies does not provide an unambiguous measure of brain function; however, we hypothesized that central suppression or attenuation of somatosensation on the moving limb would likely lead to a relative reduction in BOLD signal in brain areas associated with somatosensation. Behavioral data collected in the MRI scanner were analyzed as before. The proportion of “after” responses was calculated for each participant at each SMI for each limb and a logistic function was fitted to these data. Note that all somatosensory stimuli were substantially above threshold and easily detected by all individuals. The mean PSS values from the individually fitted logistic functions for the moving and nonmoving limbs were: moving arm = −53 (± 103)
ms; nonmoving arm = 2 (± 64) ms. This difference was tested using a nonparametric permutation test which confirmed that the difference in median values of −32.6 ms was statistically significant (p < .05) (see Supplementary material for details). These data again confirm that, consistent with FM accounts, participants judged that somatosensory stimuli delivered to their moving limb occurred later than when identical stimuli were delivered to the nonmoving limb. Brain imaging data were analyzed using standard techniques within BrainVoyagerQX software (see Methods). Four predictors were defined as follows: (1) stimulation of the right arm in the absence of movement; (2) stimulation of the left arm in the absence of movement, (3) stimulation to the right (moving) arm; and (4) stimulation to the left (nonmoving) arm with movement of the right arm. A random effects general linear model (GLM) analysis was carried out to analyze effects at the group level, and the contrast of somatosensory stimulation alone (Predictors 1 and 2) vs. no stimulation was used to determine four functionally defined volumes of interest (VOIs) that were activated by somatosensory stimulation in the absence of overt movements. The analyses revealed increased BOLD activation bilaterally within the parietal operculum (which includes SII; Eickhoff, Amunts, Mohlberg, & Zilles, 2006) and bilaterally
MODULATION OF SOMATOSENSORY PERCEPTION
S1
0.5
0
prop. "after" res pons es
–175 1
0
0.8 0.6 0.4 0.2 0
0
1
S2
0.5
0
175
S3
–175
–175 1
0
175
0
175
S4
0.8 0.6 0.4 0.2 0
175
–175
(B) 1
0.8 prop. "after" responses
Downloaded by [University of Birmingham] at 16:57 09 January 2015
prop. "after" res pons es
1
prop. "after" res pons es
prop. "after" res pons es
(A)
0.6
0.4
0.2
0 –175
0
175
Figure 2. (A) Illustrates psychometric functions fitted to the TOJs of four representative participants. Data from the two movement conditions are displayed as follows: = moving arm; = nonmoving arm. Brief somatosensory stimulation was delivered to either the limb being readied for movement or the alternate limb (counterbalanced across individuals). On each trial the participants reported a TOJ between the onset of their movement and the onset of the somatosensory stimulation. Psychometric functions are fitted separately for the moving and nonmoving arm. For each individual the psychometric function, and in particular the PSS (see Methods), for the moving limb is shifted leftwards, consistent with perception of the somatosensory stimulus being delayed for the moving limb. (B) Illustrates TOJ data for the entire group. Data from the two movement conditions are displayed as follows: = moving arm; = nonmoving arm. This confirms that the pattern seen in individual participants is replicated in the group data. Where the PSS for the moving arm is shifted leftwards and is more negative than the PSS for the nonmoving arm, it is the case that a stimulus must be delivered earlier to the moving arm than to the nonmoving arm for these two events to be perceived as occurring simultaneously.
51
within the insular cortex (Table 1, Figure 3) (threshold Z = 3.29, puncorrected < .001). These findings are consistent with previous brain imaging studies of somatosensation (Burton, 2001; Olausson et al., 2002). It should be noted that the arm region of SI was also activated bilaterally as expected, but only if a less conservative statistical threshold (puncorrected < .05) was adopted. Analyses of individual data suggest that this may be due to variability in the location of SI activations following stimulation of the forearm. For this reason SI responses will not be considered further. Previous studies have shown that ipsilateral SII responses depend on projections originating in contralateral SI and SII (Burton, 2001). Additional analyses confirmed that unilateral somatosensory stimulation produced bilateral BOLD activation in both the parietal operculum and insula VOIs (see Supplementary material for details). The size of the activation clusters observed in the parietal operculum make it rather difficult to precisely localize these activations to any specific subdivision of the parietal operculum (e.g., OP1, OP4); however, it is likely that our activations coincide with area PV, which has been linked to sensorimotor integration (Burton, 2001). The influence of movement preparation on somatosensory responses can be investigated most directly by contrasting, for each of the VOIs identified in the stimulation alone (no movement) condition, how movement preparation modulates BOLD response in that VOI for stimulation to delivered to the limb being prepared for movement compared to when identical stimulation is delivered to the nonmoving limb. To address this issue we ran a three-way ANOVA with the following factors: VOI (OP vs. INS); Limb (Moving vs. Nonmoving); and Hemisphere (Left vs. Right) (note that only trials with a negative SMI, i.e., stimulation was delivered before movement onset, were entered into these analyses). The ANOVA revealed that for the OP VOI there was a significant Limb by Hemisphere interaction effect, F(1, 11) = 9.7, p < .01, that was absent for the INS VOI, F(1, 11) = 0.25, p = .6). The basis of this interaction effect was that for the VOI in the right hemisphere OP, i.e. ipsilateral to the moving limb, the BOLD response to a stimulus delivered to the moving limb was significantly reduced compared to the BOLD response to a stimulus on the nonmoving limb (Figure 4A) (means: moving limb = 7.1; nonmoving limb = 9.9; t(11) = −3.1, p < .01) (Figure 4B, 4C). This effect was not apparent for the left hemisphere OP VOI, t(11) = 1.2, p > .1. Previous studies have demonstrated that the parietal operculum (OP) and insular (INS) cortex invariably
52
PARKINSON ET AL. TABLE 1 Significant BOLD activations present during somatosensory stimulation in the absence of movement
Contrast Stimulation alone
Region
Talairach (peak activation)
Z-value
Cluster size (voxels)
Left parietal operculum (SII) region Right parietal operculum (SII) region Left insular cortex Right insular cortex
−57–31 13 54 −22 16 −36 −19 10 48 −7 4
6.7 6.04 5.77 6.31
3034 1490 2022 623
Downloaded by [University of Birmingham] at 16:57 09 January 2015
Notes: Activations were initially thresholded at a value corresponding to q (FDR)
