An Emerging Target Paradigm to Evoke Fast Visuomotor Responses on Human Upper Limb Muscles

Summary

Presented here is a behavioral paradigm that elicits robust fast visuomotor responses on human upper limb muscles during visually guided reaches.

Abstract

To reach towards a seen object, visual information has to be transformed into motor commands. Visual information such as the object’s color, shape, and size are processed and integrated within numerous brain areas, then ultimately relayed to the motor periphery. In some instances, a reaction is needed as fast as possible. These fast visuomotor transformations, and their underlying neurological substrates, are poorly understood in humans as they have lacked a reliable biomarker. Stimulus-locked responses (SLRs) are short latency (<100 ms) bursts of electromyographic (EMG) activity representing the first wave of muscle recruitment influenced by visual stimulus presentation. SLRs provide a quantifiable output of rapid visuomotor transformations, but SLRs have not been consistently observed in all subjects in past studies. Here we describe a new, behavioral paradigm featuring the sudden emergence of a moving target below an obstacle that consistently evokes robust SLRs. Human participants generated visually guided reaches toward or away from the emerging target using a robotic manipulandum while surface electrodes recorded EMG activity from the pectoralis major muscle. In comparison to previous studies that investigated SLRs using static stimuli, the SLRs evoked with this emerging target paradigm were larger, evolved earlier, and were present in all participants. Reach reaction times (RTs) were also expedited in the emerging target paradigm. This paradigm affords numerous opportunities for modification that could permit systematic study of the impact of various sensory, cognitive, and motor manipulations on fast visuomotor responses. Overall, our results demonstrate that an emerging target paradigm is capable of consistently and robustly evoking activity within a fast visuomotor system.

Introduction

When we notice a message on our cellphone, we are prompted to perform a visually guided reach to pick up our phone and read the message. Visual features such as the shape and size of the phone are transformed into motor commands allowing us to successfully reach the goal. Such visuomotor transformations may be studied in laboratory conditions, which permit a high degree of control. However, there are scenarios where response time is important, e.g., catching the phone if it were to fall. Laboratory studies of fast visuomotor behaviors often rely on displaced target paradigms where on-going movements are modified in mid-flight following some change in target position (e.g., see ref.^1,2). While such online corrections can occur in <150 ms³, it is difficult to ascertain the exact timing of fast visuomotor output using kinematics alone due to the low-pass filtering characteristics of the arm, and because fast visuomotor output supersedes a movement already in mid-flight. Such complications lead to uncertainty about the substrates underlying fast visuomotor responses (see ref.⁴ for review). Some studies suggest that subcortical structures such as the superior colliculus, rather than fronto-parietal cortical areas, may initiate online corrections⁵.

This uncertainty regarding the underlying neural substrates may be due, at least in part, to the lack of a reliable biomarker for the output of the fast visuomotor system. Recently, we have described a measure of fast visuomotor responses that may be generated from static postures and recorded via electromyography (EMG). Stimulus-locked responses (SLRs) are time locked bursts of EMG activity that precede voluntary movement^6,7, evolving consistently ~100 ms after stimulus onset. As the name implies, SLRs are evoked by stimulus onset, persisting even if an eventual movement is withheld⁸ or moves in the opposite direction⁹. Furthermore, SLRs evoked by target displacement in a dynamic paradigm are associated with shorter latency online corrections¹⁰. Thus, SLRs provide an objective measure to systematically study the output of a fast visuomotor system involved in short latency RTs, as they may be generated from a static posture and parsed from other EMG signals unrelated to the initial phase of the fast visuomotor response.

The goal of the current study is to present a visually-guided reaching paradigm that robustly elicits SLRs. Previous studies investigating the SLR have reported less than 100% detection rates across participants, even when using more invasive intramuscular recordings^6,8,9. Low detection rates and a reliance on invasive recordings limit the usefulness of SLR measures in future investigations into the fast visuomotor system in disease or across the lifespan. While some subjects may simply not express SLRs, the stimuli and behavioral paradigms used previously may not have been ideal to evoke the SLR. Past reports of SLRs have typically used paradigms wherein participants generate visually-guided reaches towards static, suddenly appearing targets^6,9. However, a fast visuomotor system is the most likely needed in scenarios where one must rapidly interact with a falling or flying object, leading one to wonder if moving rather than static stimuli may better evoke SLRs. Therefore, we have adapted a moving target paradigm used to study eye movements¹¹, and combined it with a pro/anti visually guided reaching task used to examine the SLR⁹. When compared to results from paradigms used previously^6,8,9, it was found that SLRs in the emerging target paradigm evolved sooner, attained higher magnitudes, and were more prevalent across our participant sample. Overall, the emerging target paradigm promotes the expression of fast visuomotor responses to such a degree that objective EMG measures can be made reliably with surface recordings, potentiating study within clinical populations and across the lifespan. Further, the emerging target paradigm can be modified in many different ways, promoting more thorough investigations into the sensory, cognitive, and motor factors that promote or modify fast visuomotor responses.

Protocol

All procedures were approved by the Health Science Research Ethics Board at the University of Western Ontario. All participants provided informed consent, were paid for their participation, and were free to withdraw from the experiment at any time.

1. Participant preparation

NOTE: A small sample of healthy, young participants was studied (3 female, 2 male; mean age: 26 years +/- 3.5). All participants were right-handed and had normal or corrected-to-normal vision, with no current visual, neurological, or musculoskeletal disorders. Participants with a history of musculoskeletal upper limb injury or disorders were excluded.

1. Apply EMG sensors to the targeted upper limb muscle involved in the reaching movement being studied. Here, EMG recordings were made from the clavicular head of the right pectoralis major muscle, which is recruited for cross-body (leftward) reaching.
   NOTE: Recordings can be made from other muscles of the upper limb, or from the sternal or lateral portion of the pectoralis major muscle.
   1. Visualize the target muscle by requesting an action known to recruit the muscle of interest. For the clavicular head of the pectoralis major muscle, ask the participant to relax their elbows at their sides and push their palms together. If having difficulty visualizing the target muscle, palpate the area of interest while having the participant repeatedly perform the requested action, and target areas with notable changes in the muscle for electrode placement.
      NOTE: Visualization refers to the identification of target muscle, via seeing the muscle’s shape through the overlying skin as the participant performs an action that recruits the muscle. Visualization aids localization of the target muscle.
   2. Using alcohol swabs, clean the skin surface over the target muscle where the electrode will be placed, and also over the area where a ground electrode will be located.
   3. Prepare the surface sensors by applying adhesives and electrode gel.
   4. Ask the participant to perform the action associated with the muscle recruitment again, and adhere sensors over the muscle belly, positioning them to lie in parallel with the direction of the fibres of the targeted muscle. Place the ground electrode on the clavicle contralateral to the reaching arm. Secure sensors and ground electrodes to the surrounding skin with medical tape. Turn on the EMG system to allow for EMG collection throughout the experiment.
      NOTE: After the placement of the EMG electrodes, EMG data is collected passively and continuously throughout the experiment via the EMG system and saved as an analog datastream for later analysis.
   5. Check the quality of the EMG signal by using a desktop monitor or oscilloscope connected to the EMG system. To determine the suitable quality, have the participant perform a reaching movement into or opposite from the preferred direction of the muscle of interest, and ensure that EMG activity increases or decreases, respectively. If there is no activity at rest, then ensure that EMG activity does not increase for movement in the non-preferred direction.
      NOTE: Muscle signal quality from surface electrodes depends on many characteristics (e.g., idiosyncratic distribution of adipose tissue, subject posture). The peak EMG activity associated with movement in the preferred (contracting) direction is recommended to be at least 2x the level of activity at rest but should be considerably higher.
   6. Reposition the electrodes if necessary, to ensure that these activity levels are observed. Leave the viewing monitor or oscilloscope connected throughout the experiment to continuously monitor EMG output.
2. Set up the specific participant with the applied EMG sensors in a robotic reaching apparatus which allows reaching movements in a horizontal plane, and the application of force to the manipulandum.
   NOTE: Adding force against the muscle of interest increases the background activity, allowing for the expression of the SLR as an increase or decrease in muscle activity following stimulus presentation in the muscle’s preferred or non-preferred direction, respectively. A level of baseline activity is especially useful in the non-preferred direction, as baseline and non-preferred reaching activity would be indistinguishable without a background loading force. An applied force of 5N to the right and 2N of force down (opposite to a leftward presented target relative to the start position), throughout the entirety of the experiment can be sufficient. The force should remain constant throughout the experiment, so lower forces can be used if necessary.
   1. Seat the participant in the experimental chair, prioritizing participant comfort with respect to the added forced against the limb to minimize changes of posture throughout the experiment.

2. Stimuli construction/ apparatus

1. Generate all experimental procedures and stimuli in the robotic reaching apparatus with a built-in visual display.
   NOTE: Ensure that the robotic reaching apparatus is equipped with an interface between visual output and manipulandum motor output that permits simultaneous analog (e.g., manipulandum position, photodiode output) and EMG recordings. Ensure that this apparatus is equipped with software capable of running blocks of individual, pre-programmed trials with all pre-programmed visual components. The built-in visual display may be a standard monitor or customized high-quality projector; however, higher quality projectors are recommended to ensure temporal and visual resolution of the displayed target.
   1. Generate the 4 primary components of the emerging target paradigm (see Supplemental Figure 1) via built in software that drives the visual display.
      NOTE: Components should all be generated via built in software which projects the specified components onto the visual display during each data collection session. Each component is manually entered into the software, which transforms input coordinates for shapes into shapes seen on the visual display. A complete coding of all components and target motions is done prior to data collection, therefore, no experimenter intervention of the paradigm is required during the data collection, as the paradigm runs automatically based on the participant’s responses. The following coordinates (reported in cm) are referenced in relation to the midpoint of the two robot manipulandum origins in the robotic reaching apparatus used to collect data from participants in the current manuscript. All components of the paradigm are visible to the participant throughout each trial, except for the start position which disappears after appearance of the moving target. A different apparatus may use a different reference frame.
      1. Generate an inverted y path by manually inputting coordinates for six rectangles with the following coordinates (y: - 19 (top of inverted y) or -34 (bottom of inverted y), x:-/+2 (inner, bottom inverted y), -/+8 (outer bottom inverted y); width .5 height: 20 (top) or 15 (bottom)).
      2. Generate an occluder by manually inputting coordinates for one large rectangle (centered at: 0, -29; width: 35 height: 15) overlaying the center of the inverted y path. The color of this occluder may vary from trial to trial, providing an instruction to the participant.
         NOTE: The occluder contains a notch cut out on the center bottom between the two outputs (0, -29; width: 5 height: 5). The participant is instructed to: “fixate the notch while a target is behind the occluder”. Doing so ensures the eye is stable at target emergence. The occluder will either be colored red or green at the beginning of each trial.
      3. Generate a moving target by manually inputting coordinates for one circle which will eventually move down the inverted y and behind the occluder (start: 0, -17; radius: 1; speed: 10 cm/s, speed behind occluder: 30 cm/s).
         NOTE: The moving target (T1) is visible and stationary at the beginning of every trial.
      4. Generate how the target will move in the software by specifying the x and y coordinates of target motion.
         NOTE: The speed of target is calculated by the distance of the successive x and y coordinates. Proper presentation of target motion is dependent on the ability of the software and visual display to properly update each x and y position in rapid succession. In the software, change the status of the moving target to “invisible to participant” when the x and y position of the target have moved fully below the occluder until the x and y position have fully emerged from the occluder.
      5. Generate a start position (0, -42; radius 1). The participant will need to acquire this position to initiate each trial.
   2. Generate a real time cursor (RTC) representing the participant’s hand position on the screen in real time.
      NOTE: The participant’s hand/arm was occluded during the experiment via an upward facing mirror reflecting downward-presented targets. This may be done via built in software functions specific to the apparatus, which places a target overtop of the continually updated x and y coordinates of the hand.

3. Procedure

1. Click the “Start” button on the associated software presented on the experimenter’s screen, which initiates the first trial and force generated by the robotic reaching apparatus applied to the participant's upper limb.
   NOTE: After the experimenter clicks start, no intervention is required by the experimenter, until between blocks where the experimenter must press start again. Experimenter intervention may also be required if the EMG signal being continually monitored changes, or the participant is unable to complete the experiment. All experiments should be stopped immediately if an emergency arises. Force applied to the participant's hand is automatically stopped if the participant lets go of the handle via built in task programs. It is recommended that an apparatus with a button to terminate the experiment in emergency situations is used.
   1. Verbally instruct the participant to start the first trial by bringing the RTC (indicated by the position of the manipulandum) into the start position (T0) for a variable duration of 1- 1.5 s. The occluder changes color to instruct the subject that the upcoming trial requires a pro- or anti-reach.
      NOTE: Bringing RTC into T0 initiates each trial. If the participant exits the T0 start position before the prescribed time, the trial will start again once RTC is back in T0.
   2. Ensure that the moving target (T1) which was stationary and visible to the participant at the top of the inverted y (2.1.1.3), begins movement towards the participant along the path of the inverted y, which was initiated by the participant bringing RTC into T0 in the previous step.
      NOTE: When T1 starts moving, T0 disappears. No restrictions are placed on the participant’s arm after this time, however, the participant is instructed to stay within the imagined confines of T0.
   3. Ensure that T1 moves behind the occluder and is invisible to the participant. During this interval, the participant maintains hand position at imagined T0.
   4. Ensure that T1 travels behind the occluder at a constant speed of 30 cm/s along the y axis towards the participant. Once T1 reaches half the length of the occluder, it bifurcates along one of the inverted y outputs with an additional x velocity component. Thus, speed along the y axis is kept constant. The target vanishes for a constant delay of ~0.5 s, with the delay depending on the size of the occluder and the speed of T1 motion.
   5. When T1 reaches the edge of the occluder closest to the participant, ensure that the software program does not present T1 as emerging by sliding past the edge of the occluder, as doing so would initially present a “half-moon” stimulus to the visual system. Instead, check that the software program keeps T1 invisible until the full target has emerged, and then presents it to the participant.
      NOTE: This is done to control for visual processing effects of partial stimuli, especially if different speeds of targets are used which would cross the boundary at different times. A partial emergence of a target (e.g., half moon stimulus) produces a target composed initially of a higher spatial frequency, which based on previous results would lead to increased SLR latency and decreased magnitude¹⁰.
   6. Check that the software program presents T1 to a randomized side at one of the two inverted y paths while the participant’s hand remains stationary at T0.
      NOTE: Simultaneous with the emergence of T1 from below the occluder, a secondary target is presented in the corner of the screen, at a location covered by a photodiode. This target presented to the photodiode is not seen by the subject but provides an analog signal to a photodiode integrated in the robotic reaching device. This photodiode signal allows for the precise alignment of target appearance with muscle activity and ensures no lags or delays are present within the robotic reaching apparatus.
   7. When T1 emerges from the behind the occluder, see if the participant is able to generate a visually guided reach depending on the color of the occluder. When the occluder is green, ask the participant to intercept T1 with the RTC. When the occluder is red, ask the participant to move the RTC away from T1.
      NOTE: A green occluder color (2.1.1.2) indicates a pro reach (i.e., towards the occulder) and a red color indicated away from moving target T1 (i.e., an anti-reach). In the anti-reach condition, a correct interception is not based on the mirror image of T1, but rather the horizontal distance relative to T0.
   8. Depending on their reaching behavior, provide feedback as either a ‘hit’ (correct interception), ‘wrongway’ (incorrect direction for pro/anti reach), or ‘miss’ (neither correct nor incorrect responses detected) during the inter-trial interval. This feedback consists of text written on the occluder.
   9. Make sure that T1 and T0 reappear at their respective original locations 200 ms after the participant’s reach behavior is completed. Start the next trial when the participant brings the RTC to T0.
2. Ask each participant to perform 4 blocks of 100 trials, yielding 100 reaches per condition. Randomize the trial types intermixed with pro or anti-reaches after left and right stimuli. Each block takes approximately 7.5 min to complete.
   NOTE: It is recommended that each condition consists of a minimum of ~80 repeats when using surface recordings, as the next analysis step relies on data from many trials for SLR detection.
   1. Minimize the participant movement between each block to ensure consistency of recordings. After verbal confirmation that the participant is ready to begin the next block, initiate the next block and continue to monitor participant performance and EMG output.
      NOTE: Continued monitoring of EMG output via a desktop monitor by the experimenter may be required for detecting issues with surface EMG recordings. For example, during extended periods of reaching movements, surface EMG electrodes may become unstuck from the participant’s skin due to sweating.
3. Collect data from a control static paradigm to enable comparison of data to that obtained in the emerging target paradigm.
   NOTE: This may be done before or after the emerging target paradigm. To create a control static paradigm, repeat steps 2.1.1.3, 2.1.1.5, 2.2, 3.1, 3.1.1, 3.1.7, 3.2 and 3.2.1; however, in step 2.1.1.3, do not code T1 starting at the top of the screen and moving towards the participant. Instead, position T1 to appear either to the left or right of T0. Furthermore, T0 is now either red or green akin to the occluder used in the emerging target paradigm. The trial proceeds as described below.
   1. Verbally instruct the participant to bring the RTC into T0 to start the first trial, which is in the same location as in the emerging target paradigm.
   2. Make sure that the software program presents T0 as either red or green to indicate a pro or anti reach respectively. Randomize the hold period of 1-2s for the participant to hold the RTC in T0.
   3. Make sure that the software program presents a static target either to the left or right, 10 cm from T0. Randomize the target side across trials.
   4. As in the emerging target paradigm, ask the participant to reach towards a target if T0 is green, and reach in the diametrically opposite direction away from a target if T0 is red. The next trial proceeds after contact with a target or anti-target location.
   5. Ensure that each participant performed 4 blocks of 100 trials, yielding 100 reaches per condition. Trial types were randomly intermixed.

4. Analysis

1. Analyze all data in offline custom scripts and discard error trials.
   NOTE: Error trials are defined by incorrect reach directions (3.5 cm), long RTs (>500 ms) indicating presumed inattentiveness or short RTs (<120) indicating anticipation.
   1. Derive the reaction time (RT) for reaching movements for each trial by identifying the time at which movement exceeded 8% of the peak tangential velocity.
      NOTE: Other methods to define RT may be used.
   2. To analyze the muscle activity, use offline scripts to convert the EMG signals to source microvolts, remove any DC offset, rectify the EMG signal, and filter the signal with a 7-point moving average filter.
   3. Use a time-series receiver-operating characteristic (ROC) analysis to detect the presence and latency of the SLR^6,7.
      NOTE: Alternative methods for determining the time-locked nature of SLR activity may be used.
      1. To perform the time-series ROC analysis, segregate EMG data based on the side of target presentation and trial condition (Figure 1a shows left versus right data for pro-reaches).
      2. Calculate the area under the ROC curve for the two populations, for every time sample (1 ms) from 100 ms before to 300 ms after the target presentation (e.g., Figure 2c).
         NOTE: ROC value of 0.5 indicates chance discrimination, whereas values of 1 or 0 indicate perfectly correct or incorrect discrimination relative to target presentation, respectively.
      3. Determine discrimination latency as the first of 8 of 10 consecutive points that exceeded a value of 0.6 (Figure 2c indicated by vertical red or blue lines).
         NOTE: Threshold, and number of points exceeding threshold may change depending on the quality and quantity of surface or intramuscular EMG recordings, and a bootstrapping analysis may be used to objectively determine confidence intervals. Past work has shown that a value of 0.6 equates approximately to a 95% confidence interval¹².
   4. To determine the presence of an SLR on pro-reach trials, use an RT-split analysis (see Figure 1⁸), whereby steps 4.1.3.2 and 4.1.3.3 are performed separately on the early and late half of reaches based on RT (Figure 1a purple trials, and green trials).
      1. Plot the early discrimination time and mean early RT as one point, then plot late discrimination time and mean late RT as a second point on the same plot. Connect these two points with a line (Figure 1c). An SLR is detected when the slope of this line exceeds 67.5˚.
         NOTE: For this line, a slope of 90˚ would indicate that EMG discrimination times are perfectly locked to stimulus presentation (as EMG activity is initiated at the same latency, regardless of ensuing movement time), whereas a slope of 45˚ would indicate that EMG discrimination are perfectly locked to movement onset. In practice, a cut-off slope of 67.5˚ (halfway between 45˚ and 90˚) is used to detect whether an SLR was present (slope > 67.5˚) or not (slope < 67.5˚); as this indicates that EMG activity is more locked to stimulus rather than movement onset.
   5. If SLR presence is determined, define the SLR latency by the discrimination latency from all trials (4.1.3.3).
   6. Define the SLR magnitude as the difference between left and right mean EMG traces (e.g., Figure 2c dark red versus light red traces, or dark blue versus light blue traces) from SLR latency to 30 ms post discrimination latency.
      NOTE: Magnitude time values may be extended or shortened.

Results

Stimulus locked responses (SLRs) are brief bursts of muscle activity time locked to the stimulus onset that evolve well before the larger volley of muscle recruitment associated with movement onset. The time-locked nature of the SLR produced a ‘banding’ of muscle activity visible at ~100 ms when viewing all trials sorted for reaction time (RT) (Figure 1a, highlighted by grey boxes). As shown in Figure 1a, SLRs was dependent on target location, with S...

Discussion

Humans have a remarkable capacity, when needed, to generate rapid, visually guided actions at latencies that approach minimal afferent and efferent conduction delays. We have previously described stimulus-locked responses (SLRs) on the upper limb as a new measure for rapid visuomotor responses^6,9,10. While beneficial in providing a trial-by-trial benchmark for the first aspect of upper limb muscle recruitment influenced by the v...

Materials

Name	Company	Catalog Number	Comments
Bagnoli-8 Desktop Surface EMG System	Delsys Inc.		Another reaching apparatus may be used
Kinarm End-Point Robot	Kinarm, Kingston, Ontario, Canada		Another reaching apparatus may be used
MATLAB (version R2016a) Stateflow and Simulink applications	The MathWorks, Inc., Natick, Massachusetts, United States
PROPixx projector	VPIXX Saint-Bruno, QC, Canada		This is a custom built addon for the Kinarm. Other displays may be used.
			Resolution: 1920 x 1080. Standard viewing monitors may also be used.