Exp Brain Res (2015) 233:927–936 DOI 10.1007/s00221-014-4168-6
RESEARCH ARTICLE
Task‑relevancy effects on movement‑related gating are modulated by continuous theta‑burst stimulation of the dorsolateral prefrontal cortex and primary somatosensory cortex Katlyn E. Brown · Jennifer K. Ferris · Mohammad A. Amanian · W. Richard Staines · Lara A. Boyd
Received: 3 February 2014 / Accepted: 27 November 2014 / Published online: 16 December 2014 © Springer-Verlag Berlin Heidelberg 2014
Abstract Movement-related gating ensures that decreased somatosensory information from external stimulation reaches the cortex during movement when compared to resting levels; however, gating may be influenced by task-relevant manipulations, such that increased sensory information ascends to the cortex when information is relevant to goal-based actions. These task-relevancy effects are hypothesized to be controlled by a network involving the dorsolateral prefrontal cortex (DLPFC) based on this region’s known role in selective attention, modulating the primary somatosensory cortex (S1). The purpose of the current study was first to verify task-relevancy influences on movement-related gating in the upper limb, and second to test the contribution of the DLPFC and the primary somatosensory cortex (S1) to these relevancy effects. Ten healthy participants received median nerve stimulation at the left wrist during three conditions: rest, task-irrelevant movement, and task-relevant movement. Cortical responses to median nerve stimulations were measured in the form of somatosensory evoked potentials (SEPs). The three conditions
were collected on a baseline day and on two separate days following continuous theta-burst (cTBS), which transiently reduces cortical excitability, over either the contralateral S1 or DLPFC. Results demonstrated a significant interaction between stimulation and condition, with a priori contrasts revealing that cTBS over either S1 or DLPFC diminished the relevancy-based modulation of SEP amplitudes; however, the degree of this effect was different. These results indicate that DLPFC influences over S1 are involved in the facilitation of relevant sensory information during movement.
K. E. Brown · L. A. Boyd (*) Department of Rehabilitation Science, Faculty of Medicine, University of British Columbia, 212‑2177 Wesbrook Mall, Vancouver, BC V6T 1Z3, Canada e-mail: [email protected]
W. R. Staines Department of Kinesiology, Faculty of Applied Health Sciences, University of Waterloo, 200 University Ave W, Waterloo, ON N2L 3G1, Canada
K. E. Brown e-mail: [email protected] J. K. Ferris Department of Neuroscience, Faculty of Medicine, University of British Columbia, 2255 Wesbrook Mall, Vancouver, BC V6T 1Z3, Canada
Keywords Somatosensory · Attention · Task-relevant · Prefrontal cortex · Gating
Introduction Successful interaction with one’s environment depends on the ability to extract relevant sensory information amidst
L. A. Boyd Brain Research Centre, University of British Columbia, 2211 Wesbrook Mall, Vancouver, BC V6T 2B5, Canada L. A. Boyd Department of Physical Therapy, Faculty of Medicine, University of British Columbia, 212‑2177 Wesbrook Mall, Vancouver, BC V6T 1Z3, Canada
M. A. Amanian Department of Electrical and Computer Engineering, Faculty of Applied Science, University of British Columbia, 2332 Main Mall, Vancouver, BC V6T 1Z4, Canada
13
928
an overwhelming barrage of concurrent sensory feedback. During movement, sensory information from the periphery is filtered as it ascends to the cortex; the most important, or relevant, information causes the greatest response in the primary somatosensory cortex (S1) in healthy individuals (Staines et al. 2000, 2002a). Past work documented the neural mechanisms underlying the ability to filter incoming sensory information during movement in a phenomenon referred to as “movement-related gating”. Early somatosensory evoked potential (SEP) components generated in S1 (i.e. N19–P25 for median nerve SEPs) have reduced amplitude during both active and passive movements (Rushton et al. 1981; Abbruzzese et al. 1981). Cortically, descending motor output from the primary motor cortex (M1) is thought to influence incoming sensory information in several ways: at the level of the spinal cord by altering levels of primary afferent depolarization (PAD) (Seki et al. 2009), at the level of the dorsal column nuclei by way of presynaptic inhibition (Canedo 1997), and thalamically through interconnections with the thalamic reticular nucleus (TRN) and ventral posterolateral nucleus (VPL) (Guillery et al. 1998). The observation of gated SEP levels in S1 80–100 ms prior to movement provides support for cortical contributions to gating (Jones et al. 1989; Cohen and Starr 1987). Gating is also seen during passive movement indicating a role for peripheral or spinal mechanisms (Jones et al. 1989). Staines et al. (2000) showed that task-relevancy influences somatosensory information processing and sensory gating in the lower limbs. This study contained three major components: a rest condition, a task-irrelevant movement condition, and a task-relevant movement condition (Staines et al. 2000). In the task-irrelevant condition, participants’ feet were passively moved through a series of plantar flexion and dorsiflexion movements with no focus on proprioceptive feedback; in the task-relevant condition, participants were instructed to attend to the position of their passively moved foot and match that position with their other foot following the episodic passive movement, engaging the proprioceptive system. Nerve stimulation in the rest condition produced the largest cortical responses, indexed by early SEP components indicative of the arrival of afferent information in S1. Both the task-relevant and task-irrelevant conditions produced lower early SEP components as compared to the rest condition (Staines et al. 2000). Responses to stimuli in S1 were also modulated by the relevancy assigned to the stimuli by the individual; that is, sensory information that was more closely attended to by the subject evoked a larger response in S1 than irrelevant sensory information (Staines et al. 2000, 2002a). While the mechanisms underlying movement-related gating have been well documented, the neural network underpinning the interaction between attentional systems and
13
Exp Brain Res (2015) 233:927–936
movement-related gating is less well understood. The dorsolateral prefrontal cortex (DLPFC) is a likely site of control for the effects of task-relevancy on movement-related gating. The DLPFC has been linked to both attentional and somatosensory processing (Staines et al. 2000; Knight et al. 1999; Bolton and Staines 2011), and following damage to the DLPFC attention deficits are common (Hurford et al. 2013; Yamaguchi and Knight 1990). Specifically, prefrontal damage may result in an inability to filter out irrelevant incoming sensory stimuli, impairing the ability to attend to specific sensory stimuli in the presence of distractors (Knight et al. 1999). To examine this network in a healthy brain, a recent study investigated the effect of transient inhibition of DLPFC on a tactile attention task (Bolton and Staines 2011). The study employed continuous theta-burst stimulation (cTBS), which transiently reduces excitability in a localized cortical area (Huang et al. 2005), to temporarily disrupt contralateral DLPFC. Individuals performed a tactile oddball task in which they were to detect tactile “target” stimuli on a predetermined digit, which occurred 25 % of the time and were smaller than the regularly occurring stimuli. Stimulation was administered to two digits, but individuals were instructed to attend only to one; this allowed stimuli to be divided into relevant and irrelevant groups (Bolton and Staines 2011). The P100 potential, measured via electroencephalography (EEG), from the contralateral secondary somatosensory cortex (SII) is typically inhibited when stimuli are irrelevant; however, following cTBS over DLPFC, this inhibition was not seen. This suggests that the DLPFC is an integral part of somatosensory-based attentional networks (Bolton and Staines 2011). Further supporting a role for the DLPFC in sensory processing, functional magnetic resonance imaging (fMRI) data show that when healthy controls were instructed to attend only to relevant tactile information, increased cortical activity was noted in both contralateral S1 and DLPFC; at the same time, activation decreased in ipsilateral S1 (Staines et al. 2002b). When the tactile stimuli were irrelevant, increased activity in S1 was not evident, but DLPFC activity remained elevated (Staines et al. 2002a, b), implying a role for attentional processes in modulating S1 activity. Moreover, parallel task-related changes in DLPFC activity suggest a prefrontal cortical gating mechanism (Staines et al. 2002b). Thus, a network involving prefrontal influence on S1 may be responsible for relevancy-based filtering of sensory information (Staines et al. 2002a, b). The purpose of the current study was first to establish whether task-relevancy effects, which have previously only been shown in the lower limb (Staines et al. 2000), are also present during arm movements. We hypothesized that similar to the lower limb, early SEP component (N19–P25) amplitudes would be the largest when assessed at rest, with movement causing a gating effect that would result in
Exp Brain Res (2015) 233:927–936
diminished N19–P25 amplitudes. Additionally, we predicted that when proprioceptive feedback from the movement was relevant to a task, the N19–P25 amplitudes would be higher than when the feedback was irrelevant, but would not reach resting levels. Secondly, we examined a network hypothesized to be involved in task-relevancy effects on movementrelated gating. Specifically, the roles of the DLPFC and S1 in the modulation of cortical excitability during the manipulation of task-relevancy were investigated. We hypothesized that the DLPFC would be largely responsible for observed relevancy-based facilitation. Specifically, following cTBS over DLPFC, we expected that the N19–P25 amplitudes would be the same regardless of the relevancy of the feedback and that movement conditions would both be reduced as compared to resting amplitudes. Finally, we hypothesized that, while the overall amplitudes of the SEP components may be decreased, cTBS over S1 would not have any effects on the task-relevancy aspects of movement-related gating.
Methods Participants Ten (four males, six females) healthy right-handed individuals (26 ± 5 years) participated in the current experiment. Written consent in accordance with the University of British Columbia Ethical Review Board was obtained prior to participation, and individuals were screened for contraindications to transcranial magnetic stimulation (TMS) or magnetic resonance imaging (MRI) procedures. Briefly, participants were excluded if they had any neurodegenerative or musculoskeletal disorders, as well as a history of epilepsy or head trauma. Each participant completed four sessions: (1) MRI scan, (2) baseline measurement, (3) cTBS over DLPFC, and (4) cTBS over S1. Relevancy‑based modulations of movement‑related gating in the upper limb The impact of attentional manipulations on movementrelated gating in the upper limb was explored in ten individuals. To facilitate stereotaxic guidance of non-invasive brain stimulation, each participant first underwent a T1-weighted anatomic MRI scan at the UBC 3T Research MRI Centre. Following the MRI, participants completed the experimental protocol in the Brain Behaviour Laboratory at the University of British Columbia. Dependent measures To assess the arrival of information in S1 (Brodmann areas 3b and 1), EEG was used to examine the N19–P25
929
components of SEPs resulting from median nerve stimulation. A bar electrode was placed on the left wrist, just above the median nerve. A square wave pulse (0.5 ms) was delivered (Grass SD9 Stimulator with SIU-V Isolation Unit, West Warwick, RI, USA) at an intensity of motor threshold [the intensity required to evoke a just visible twitch in the left abductor pollicis brevis (APB)]. The interstimulus interval was randomly generated between 500 and 1,000 ms to ensure that stimulation onset was not predictable. A total of 110 stimulations were delivered in each condition, and the analysis programme produced average traces. To ensure changes in SEP amplitude were the direct result of task manipulations, rather than changes in recruitment of largediameter afferent fibres, resulting from positional changes, motor wave (M-wave) amplitude was monitored. M-wave amplitude was monitored using Ag–Cl electrodes placed over the muscle belly of APB. As M-waves result from efferent nerve stimulation, this monitoring ensures a constant number of recruited muscle fibres. EEG was recorded from electrodes C4 and CP4 (contralateral sensorimotor regions) throughout each condition with a TMS-compatible cap referenced to AFz (2,000 Hz sampling rate) (NeuroPrax; Neuroconn, Ilmenau, Germany). Channel impedances were 200 µV is elicited on five out of ten trials while participants maintain an APB contraction of 20 % of their maximal force, as measured by surface EMG (Rothwell et al. 1999). cTBS consists of a
Exp Brain Res (2015) 233:927–936
931
Fig. 2 Brainsight™ images demonstrating localization of primary motor cortex (M1), primary somatosensory cortex (S1), and dorsolateral prefrontal cortex (DLPFC) in a representative participant. a
Depicts all of the stimulation sites, b shows only DLPFC, c exclusively displays M1 localization, and S1 is shown in (d)
series of bursts of three pulses 20 ms apart, repeated every 200 ms for 40 s (600 total pulses) at 80 % of AMT for APB (Huang et al. 2005). When applied over DLPFC, the coil was held at a 90° angle to the mid-sagittal line (Grossheinrich et al. 2009; Bolton and Staines 2011). When applied over S1, the coil was held at a 45° angle to the mid-sagittal line with the handle in a posterior lateral orientation (Vidoni et al. 2010). Prior to completion of any TMS measures, a highresolution anatomical MRI was collected for each participant (TR = 12.4 ms, TE = 5.4 ms, flip angle h = 35°, FOV = 256 mm, 170 slices, 1 mm thickness) at the UBC MRI Research Centre on a Phillips Achieva 3.0T wholebody MRI scanner (Phillips Healthcare, Andover, MD) using a sensitivity-encoding head coil (SENSE). The images acquired in this session were imported into BrainSight (Rogue Research Inc., Montreal, QC), a TMS neuronavigation software, in order to register the coil in stereotaxic space (Fig. 2). This registration enabled insertion of the coordinates for DLPFC into the software and ensured the TMS coil was placed over these precise coordinates. For
the DLPFC, coordinates from past functional MRI work were used (MNI: x, y, z = 40, 21, 27) (Ott et al. 2011). In order to convert the images into MNI space, the anterior and posterior commissures were identified in each individual’s MRI, and bounding boxes were set around the cortex. Using Brainsight software, the images were then converted into MNI space. To locate S1, the coil was marked on Brainsight as 2 cm posterior and 1 cm lateral to the M1 APB hotspot. To ensure stimulation of M1 was not taking place, single pulses were delivered over this location to confirm no EMG activity was recorded. The order of the sessions in which stimulation was delivered over either S1 or DLPFC was randomized across participants (as assigned by a random generation computer program). Each session was completed within 40 min after the administration of cTBS, ensuring the effect of cTBS remained for the duration of collection. The sessions occurred at least a week apart in order to ensure that stimulation of the first region would not affect stimulation of the second region (Rothwell et al. 1999). All stimulation parameters were within published safety standards (Wassermann 1998).
13
932
Exp Brain Res (2015) 233:927–936
Dependent measures The same conditions and procedures as previously described were repeated after cTBS had been applied. The participants began the session with cTBS over one of the two locations and then completed the three conditions (rest, task-irrelevant, task-relevant). EEG was collected throughout the conditions, but not during the application of TMS. The order of these conditions was once again randomized. As the same individuals participated throughout, the data used to initially replicate the taskrelevancy effects in the upper limb served as a baseline measure. Statistics A stimulation site (none, DLPFC, S1) × Condition (rest, task-irrelevant, task-relevant) repeated-measures ANOVA (dependent variable: N19–P25 amplitude) was used to evaluate the relationship between cortical stimulation and movement conditions. An interaction effect indicated that, as hypothesized, task-relevancy effects on gating are being influenced differently based on cTBS stimulation site. Planned contrasts were used to determine the locus of the interaction. Behavioural sub‑experiment A sub-experiment was collected in order to provide behavioural data and support our argument that individuals were able to attend to and successfully perform the task. This attentional component is important to ensure the conditions are, as we expect, requiring different levels of attention. In this experiment we had eight individuals (three males, five females, average age 27 ± 4 years), similar to individuals in the original experiment, perform the taskrelevant condition for ten trials. The same instructions were provided as in the original experiment, such that individuals were instructed to attend to the passive movement of their left hand, and to actively match this movement with their right hand once the passive movement had ceased and their hand was placed into a neutral position. Participants were further instructed to use homologous muscles such that if their left wrist was extended, they extended their right wrist. Measures from potentiometers in each hand device were exported into Excel. Traces were graphically represented, and peaks were extracted using a custom MatLab program. The difference in these peaks was calculated, and an individual average was determined over the ten trials. A onesample t test was used to compare these values to 0, as that would represent no difference between the two hands (perfect performance).
13
Fig. 3 Average EEG traces. a Represents baseline traces, b represents traces following cTBS over DLPFC, and c represents traces following cTBS over S1. The rest condition is shown in black, the taskirrelevant condition is shown as a dotted line, and the task-relevant condition is represented as a dashed line
Results Group average SEP figures for all conditions across all stimulation sites can be seen in Fig. 3. Relevancy‑based modulations of movement‑related gating in the upper limb A one-way repeated-measures ANOVA (dependent variable: N19–P25 amplitude) revealed a significant effect of condition (F(2,18) = 16.027, p
RESEARCH ARTICLE
Task‑relevancy effects on movement‑related gating are modulated by continuous theta‑burst stimulation of the dorsolateral prefrontal cortex and primary somatosensory cortex Katlyn E. Brown · Jennifer K. Ferris · Mohammad A. Amanian · W. Richard Staines · Lara A. Boyd
Received: 3 February 2014 / Accepted: 27 November 2014 / Published online: 16 December 2014 © Springer-Verlag Berlin Heidelberg 2014
Abstract Movement-related gating ensures that decreased somatosensory information from external stimulation reaches the cortex during movement when compared to resting levels; however, gating may be influenced by task-relevant manipulations, such that increased sensory information ascends to the cortex when information is relevant to goal-based actions. These task-relevancy effects are hypothesized to be controlled by a network involving the dorsolateral prefrontal cortex (DLPFC) based on this region’s known role in selective attention, modulating the primary somatosensory cortex (S1). The purpose of the current study was first to verify task-relevancy influences on movement-related gating in the upper limb, and second to test the contribution of the DLPFC and the primary somatosensory cortex (S1) to these relevancy effects. Ten healthy participants received median nerve stimulation at the left wrist during three conditions: rest, task-irrelevant movement, and task-relevant movement. Cortical responses to median nerve stimulations were measured in the form of somatosensory evoked potentials (SEPs). The three conditions
were collected on a baseline day and on two separate days following continuous theta-burst (cTBS), which transiently reduces cortical excitability, over either the contralateral S1 or DLPFC. Results demonstrated a significant interaction between stimulation and condition, with a priori contrasts revealing that cTBS over either S1 or DLPFC diminished the relevancy-based modulation of SEP amplitudes; however, the degree of this effect was different. These results indicate that DLPFC influences over S1 are involved in the facilitation of relevant sensory information during movement.
K. E. Brown · L. A. Boyd (*) Department of Rehabilitation Science, Faculty of Medicine, University of British Columbia, 212‑2177 Wesbrook Mall, Vancouver, BC V6T 1Z3, Canada e-mail: [email protected]
W. R. Staines Department of Kinesiology, Faculty of Applied Health Sciences, University of Waterloo, 200 University Ave W, Waterloo, ON N2L 3G1, Canada
K. E. Brown e-mail: [email protected] J. K. Ferris Department of Neuroscience, Faculty of Medicine, University of British Columbia, 2255 Wesbrook Mall, Vancouver, BC V6T 1Z3, Canada
Keywords Somatosensory · Attention · Task-relevant · Prefrontal cortex · Gating
Introduction Successful interaction with one’s environment depends on the ability to extract relevant sensory information amidst
L. A. Boyd Brain Research Centre, University of British Columbia, 2211 Wesbrook Mall, Vancouver, BC V6T 2B5, Canada L. A. Boyd Department of Physical Therapy, Faculty of Medicine, University of British Columbia, 212‑2177 Wesbrook Mall, Vancouver, BC V6T 1Z3, Canada
M. A. Amanian Department of Electrical and Computer Engineering, Faculty of Applied Science, University of British Columbia, 2332 Main Mall, Vancouver, BC V6T 1Z4, Canada
13
928
an overwhelming barrage of concurrent sensory feedback. During movement, sensory information from the periphery is filtered as it ascends to the cortex; the most important, or relevant, information causes the greatest response in the primary somatosensory cortex (S1) in healthy individuals (Staines et al. 2000, 2002a). Past work documented the neural mechanisms underlying the ability to filter incoming sensory information during movement in a phenomenon referred to as “movement-related gating”. Early somatosensory evoked potential (SEP) components generated in S1 (i.e. N19–P25 for median nerve SEPs) have reduced amplitude during both active and passive movements (Rushton et al. 1981; Abbruzzese et al. 1981). Cortically, descending motor output from the primary motor cortex (M1) is thought to influence incoming sensory information in several ways: at the level of the spinal cord by altering levels of primary afferent depolarization (PAD) (Seki et al. 2009), at the level of the dorsal column nuclei by way of presynaptic inhibition (Canedo 1997), and thalamically through interconnections with the thalamic reticular nucleus (TRN) and ventral posterolateral nucleus (VPL) (Guillery et al. 1998). The observation of gated SEP levels in S1 80–100 ms prior to movement provides support for cortical contributions to gating (Jones et al. 1989; Cohen and Starr 1987). Gating is also seen during passive movement indicating a role for peripheral or spinal mechanisms (Jones et al. 1989). Staines et al. (2000) showed that task-relevancy influences somatosensory information processing and sensory gating in the lower limbs. This study contained three major components: a rest condition, a task-irrelevant movement condition, and a task-relevant movement condition (Staines et al. 2000). In the task-irrelevant condition, participants’ feet were passively moved through a series of plantar flexion and dorsiflexion movements with no focus on proprioceptive feedback; in the task-relevant condition, participants were instructed to attend to the position of their passively moved foot and match that position with their other foot following the episodic passive movement, engaging the proprioceptive system. Nerve stimulation in the rest condition produced the largest cortical responses, indexed by early SEP components indicative of the arrival of afferent information in S1. Both the task-relevant and task-irrelevant conditions produced lower early SEP components as compared to the rest condition (Staines et al. 2000). Responses to stimuli in S1 were also modulated by the relevancy assigned to the stimuli by the individual; that is, sensory information that was more closely attended to by the subject evoked a larger response in S1 than irrelevant sensory information (Staines et al. 2000, 2002a). While the mechanisms underlying movement-related gating have been well documented, the neural network underpinning the interaction between attentional systems and
13
Exp Brain Res (2015) 233:927–936
movement-related gating is less well understood. The dorsolateral prefrontal cortex (DLPFC) is a likely site of control for the effects of task-relevancy on movement-related gating. The DLPFC has been linked to both attentional and somatosensory processing (Staines et al. 2000; Knight et al. 1999; Bolton and Staines 2011), and following damage to the DLPFC attention deficits are common (Hurford et al. 2013; Yamaguchi and Knight 1990). Specifically, prefrontal damage may result in an inability to filter out irrelevant incoming sensory stimuli, impairing the ability to attend to specific sensory stimuli in the presence of distractors (Knight et al. 1999). To examine this network in a healthy brain, a recent study investigated the effect of transient inhibition of DLPFC on a tactile attention task (Bolton and Staines 2011). The study employed continuous theta-burst stimulation (cTBS), which transiently reduces excitability in a localized cortical area (Huang et al. 2005), to temporarily disrupt contralateral DLPFC. Individuals performed a tactile oddball task in which they were to detect tactile “target” stimuli on a predetermined digit, which occurred 25 % of the time and were smaller than the regularly occurring stimuli. Stimulation was administered to two digits, but individuals were instructed to attend only to one; this allowed stimuli to be divided into relevant and irrelevant groups (Bolton and Staines 2011). The P100 potential, measured via electroencephalography (EEG), from the contralateral secondary somatosensory cortex (SII) is typically inhibited when stimuli are irrelevant; however, following cTBS over DLPFC, this inhibition was not seen. This suggests that the DLPFC is an integral part of somatosensory-based attentional networks (Bolton and Staines 2011). Further supporting a role for the DLPFC in sensory processing, functional magnetic resonance imaging (fMRI) data show that when healthy controls were instructed to attend only to relevant tactile information, increased cortical activity was noted in both contralateral S1 and DLPFC; at the same time, activation decreased in ipsilateral S1 (Staines et al. 2002b). When the tactile stimuli were irrelevant, increased activity in S1 was not evident, but DLPFC activity remained elevated (Staines et al. 2002a, b), implying a role for attentional processes in modulating S1 activity. Moreover, parallel task-related changes in DLPFC activity suggest a prefrontal cortical gating mechanism (Staines et al. 2002b). Thus, a network involving prefrontal influence on S1 may be responsible for relevancy-based filtering of sensory information (Staines et al. 2002a, b). The purpose of the current study was first to establish whether task-relevancy effects, which have previously only been shown in the lower limb (Staines et al. 2000), are also present during arm movements. We hypothesized that similar to the lower limb, early SEP component (N19–P25) amplitudes would be the largest when assessed at rest, with movement causing a gating effect that would result in
Exp Brain Res (2015) 233:927–936
diminished N19–P25 amplitudes. Additionally, we predicted that when proprioceptive feedback from the movement was relevant to a task, the N19–P25 amplitudes would be higher than when the feedback was irrelevant, but would not reach resting levels. Secondly, we examined a network hypothesized to be involved in task-relevancy effects on movementrelated gating. Specifically, the roles of the DLPFC and S1 in the modulation of cortical excitability during the manipulation of task-relevancy were investigated. We hypothesized that the DLPFC would be largely responsible for observed relevancy-based facilitation. Specifically, following cTBS over DLPFC, we expected that the N19–P25 amplitudes would be the same regardless of the relevancy of the feedback and that movement conditions would both be reduced as compared to resting amplitudes. Finally, we hypothesized that, while the overall amplitudes of the SEP components may be decreased, cTBS over S1 would not have any effects on the task-relevancy aspects of movement-related gating.
Methods Participants Ten (four males, six females) healthy right-handed individuals (26 ± 5 years) participated in the current experiment. Written consent in accordance with the University of British Columbia Ethical Review Board was obtained prior to participation, and individuals were screened for contraindications to transcranial magnetic stimulation (TMS) or magnetic resonance imaging (MRI) procedures. Briefly, participants were excluded if they had any neurodegenerative or musculoskeletal disorders, as well as a history of epilepsy or head trauma. Each participant completed four sessions: (1) MRI scan, (2) baseline measurement, (3) cTBS over DLPFC, and (4) cTBS over S1. Relevancy‑based modulations of movement‑related gating in the upper limb The impact of attentional manipulations on movementrelated gating in the upper limb was explored in ten individuals. To facilitate stereotaxic guidance of non-invasive brain stimulation, each participant first underwent a T1-weighted anatomic MRI scan at the UBC 3T Research MRI Centre. Following the MRI, participants completed the experimental protocol in the Brain Behaviour Laboratory at the University of British Columbia. Dependent measures To assess the arrival of information in S1 (Brodmann areas 3b and 1), EEG was used to examine the N19–P25
929
components of SEPs resulting from median nerve stimulation. A bar electrode was placed on the left wrist, just above the median nerve. A square wave pulse (0.5 ms) was delivered (Grass SD9 Stimulator with SIU-V Isolation Unit, West Warwick, RI, USA) at an intensity of motor threshold [the intensity required to evoke a just visible twitch in the left abductor pollicis brevis (APB)]. The interstimulus interval was randomly generated between 500 and 1,000 ms to ensure that stimulation onset was not predictable. A total of 110 stimulations were delivered in each condition, and the analysis programme produced average traces. To ensure changes in SEP amplitude were the direct result of task manipulations, rather than changes in recruitment of largediameter afferent fibres, resulting from positional changes, motor wave (M-wave) amplitude was monitored. M-wave amplitude was monitored using Ag–Cl electrodes placed over the muscle belly of APB. As M-waves result from efferent nerve stimulation, this monitoring ensures a constant number of recruited muscle fibres. EEG was recorded from electrodes C4 and CP4 (contralateral sensorimotor regions) throughout each condition with a TMS-compatible cap referenced to AFz (2,000 Hz sampling rate) (NeuroPrax; Neuroconn, Ilmenau, Germany). Channel impedances were 200 µV is elicited on five out of ten trials while participants maintain an APB contraction of 20 % of their maximal force, as measured by surface EMG (Rothwell et al. 1999). cTBS consists of a
Exp Brain Res (2015) 233:927–936
931
Fig. 2 Brainsight™ images demonstrating localization of primary motor cortex (M1), primary somatosensory cortex (S1), and dorsolateral prefrontal cortex (DLPFC) in a representative participant. a
Depicts all of the stimulation sites, b shows only DLPFC, c exclusively displays M1 localization, and S1 is shown in (d)
series of bursts of three pulses 20 ms apart, repeated every 200 ms for 40 s (600 total pulses) at 80 % of AMT for APB (Huang et al. 2005). When applied over DLPFC, the coil was held at a 90° angle to the mid-sagittal line (Grossheinrich et al. 2009; Bolton and Staines 2011). When applied over S1, the coil was held at a 45° angle to the mid-sagittal line with the handle in a posterior lateral orientation (Vidoni et al. 2010). Prior to completion of any TMS measures, a highresolution anatomical MRI was collected for each participant (TR = 12.4 ms, TE = 5.4 ms, flip angle h = 35°, FOV = 256 mm, 170 slices, 1 mm thickness) at the UBC MRI Research Centre on a Phillips Achieva 3.0T wholebody MRI scanner (Phillips Healthcare, Andover, MD) using a sensitivity-encoding head coil (SENSE). The images acquired in this session were imported into BrainSight (Rogue Research Inc., Montreal, QC), a TMS neuronavigation software, in order to register the coil in stereotaxic space (Fig. 2). This registration enabled insertion of the coordinates for DLPFC into the software and ensured the TMS coil was placed over these precise coordinates. For
the DLPFC, coordinates from past functional MRI work were used (MNI: x, y, z = 40, 21, 27) (Ott et al. 2011). In order to convert the images into MNI space, the anterior and posterior commissures were identified in each individual’s MRI, and bounding boxes were set around the cortex. Using Brainsight software, the images were then converted into MNI space. To locate S1, the coil was marked on Brainsight as 2 cm posterior and 1 cm lateral to the M1 APB hotspot. To ensure stimulation of M1 was not taking place, single pulses were delivered over this location to confirm no EMG activity was recorded. The order of the sessions in which stimulation was delivered over either S1 or DLPFC was randomized across participants (as assigned by a random generation computer program). Each session was completed within 40 min after the administration of cTBS, ensuring the effect of cTBS remained for the duration of collection. The sessions occurred at least a week apart in order to ensure that stimulation of the first region would not affect stimulation of the second region (Rothwell et al. 1999). All stimulation parameters were within published safety standards (Wassermann 1998).
13
932
Exp Brain Res (2015) 233:927–936
Dependent measures The same conditions and procedures as previously described were repeated after cTBS had been applied. The participants began the session with cTBS over one of the two locations and then completed the three conditions (rest, task-irrelevant, task-relevant). EEG was collected throughout the conditions, but not during the application of TMS. The order of these conditions was once again randomized. As the same individuals participated throughout, the data used to initially replicate the taskrelevancy effects in the upper limb served as a baseline measure. Statistics A stimulation site (none, DLPFC, S1) × Condition (rest, task-irrelevant, task-relevant) repeated-measures ANOVA (dependent variable: N19–P25 amplitude) was used to evaluate the relationship between cortical stimulation and movement conditions. An interaction effect indicated that, as hypothesized, task-relevancy effects on gating are being influenced differently based on cTBS stimulation site. Planned contrasts were used to determine the locus of the interaction. Behavioural sub‑experiment A sub-experiment was collected in order to provide behavioural data and support our argument that individuals were able to attend to and successfully perform the task. This attentional component is important to ensure the conditions are, as we expect, requiring different levels of attention. In this experiment we had eight individuals (three males, five females, average age 27 ± 4 years), similar to individuals in the original experiment, perform the taskrelevant condition for ten trials. The same instructions were provided as in the original experiment, such that individuals were instructed to attend to the passive movement of their left hand, and to actively match this movement with their right hand once the passive movement had ceased and their hand was placed into a neutral position. Participants were further instructed to use homologous muscles such that if their left wrist was extended, they extended their right wrist. Measures from potentiometers in each hand device were exported into Excel. Traces were graphically represented, and peaks were extracted using a custom MatLab program. The difference in these peaks was calculated, and an individual average was determined over the ten trials. A onesample t test was used to compare these values to 0, as that would represent no difference between the two hands (perfect performance).
13
Fig. 3 Average EEG traces. a Represents baseline traces, b represents traces following cTBS over DLPFC, and c represents traces following cTBS over S1. The rest condition is shown in black, the taskirrelevant condition is shown as a dotted line, and the task-relevant condition is represented as a dashed line
Results Group average SEP figures for all conditions across all stimulation sites can be seen in Fig. 3. Relevancy‑based modulations of movement‑related gating in the upper limb A one-way repeated-measures ANOVA (dependent variable: N19–P25 amplitude) revealed a significant effect of condition (F(2,18) = 16.027, p
