Effects of Prism Adaptation Combined with Electrical Stimulation on Cortical Excitability of Upper and Lower Limb Muscles

Abstract

Right-shifting prism adaptation (PA) is a well-characterized sensorimotor behavioral phenomenon that induces leftward visuospatial-motor behavioral aftereffects. Somatosensory electrical stimulation (Stim) has been shown to upregulate corticomotor excitability during motor training, and can accentuate the effects of PA. In this study, we investigated the effects of combining PA+Stim on corticomotor excitability and upper limb visuospatial-motor behavior, and whether these effects transfer to the lower limb. We used a single-blind, 2-arm randomized crossover repeated-measures design to compare the effects of 1 session of PA+Stim versus PA+Sham in 15 young able-bodied individuals (18-35 years). During PA, either sensory-level Stim or Sham was delivered to the left upper limb. In addition to upper limb pointing performance, corticomotor excitability, and intra-cortical inhibition were measured before and after PA in the bilateral first dorsal interossei (FDI) and left tibialis anterior and soleus muscles using motor-evoked potentials (MEPs) elicited in response to single and paired-pulse transcranial magnetic stimulation (TMS). Our results showed that behaviorally, both PA+Stim and PA+Sham showed significant sensorimotor aftereffects, inducing the expected leftward shift during upper limb pointing. Neurophysiologically, only PA+Stim, but not PA+Sham, significantly increased MEP amplitudes in the left FDI. Additionally, PA+Sham versus PA+Stim showed significant differences in PA-induced modulation of intra-cortical inhibition in the left FDI and tibialis anterior. In summary, while both PA+Stim and PA+Sham had similar behavioral aftereffects, only PA+Stim increased cortical excitability and modulated intra-cortical inhibition of the left upper and lower limb. These findings suggest that combining PA with Stim may elicit greater neurophysiological effects and inter-effector transfer to the lower limb compared to PA alone.

1. Introduction

Visual, spatial, sensory, and motor cortical circuits integrate multi-sensory inputs with motor commands to enable movement planning and execution during functional tasks such as reaching for a coffee cup or walking (Bai et al., 2021). Prism adaptation (PA) is a well-studied sensorimotor behavioral phenomenon that can be used to assess visuospatial-motor adaptation capacity. Multiple sessions of PA are also used as an intervention to modulate higher spatial cognitive processes (Boukrina & Chen, 2021; Rossetti et al., 1998) and improve visuomotor function in people with spatial neglect. Right-shifting PA involves wearing prism lenses that shift visuospatial-motor behavior towards the right hemispace, resulting in a perceptual and motor bias toward the left hemispace once the prisms are removed. The PA protocol consists of three phases. In the first phase, before prisms are applied, participants point with their upper limb at visual targets to obtain reference values (pre-adaptation pointing performance) (Farnè et al., 2002; Luauté, Halligan, et al., 2006). In the second phase, (prism exposure), participants wear right shifting glasses that induce a rightward displacement of the visual field. During early adaptation, participants initially incorrectly point to the right of the visual targets (due to a rightward shift of the target’s image); sensory feedback makes this error immediately apparent. To correct this ‘error’, participants point leftward until correct pointing to the target object is perceived, which occurs by the late adaptation period. In the third phase, after PA, when the prism glasses are removed, participants continue to over-correct and aim towards the left side of visual targets due to the preceding sensorimotor adaptation, which is measured as the prism aftereffect (Azouvi et al., 2017; Barrett et al., 2012; Magnani et al., 2014; Panico et al., 2020; Pisella et al., 2006; Prablanc et al., 2020; Rossetti et al., 1998).

While the behavioral effects of PA are well documented (Azouvi et al., 2017; Fortis et al., 2010; Rossetti et al., 1998), neural processes underlying PA need more investigation. Visuospatial information processing involves the input stage comprising perception and attention, the representational stage comprising sensorimotor integration, and the motor output stage involving intentional movements (Fortis et al., 2011). Different theories about how PA affects the visuospatial cognitive processing are debated. PA is thought to recalibrate spatial representations and strengthen the connection between perception and motor actions, thereby influencing both lower-level sensorimotor behavior as well as higher cognitive processes such as spatial awareness and attention (Rode et al., 2017; Rossetti et al., 1998). However, other studies suggest that PA may primarily affect the motor-output stage (Chaudhari et al., 2015; Fortis et al., 2011; Goedert et al., 2015). Furthermore, if and to what extent PA influences cortical motor circuits is unclear. For instance, a previous study showed that right-shifting PA modulates right hemispheric intracortical facilitation in the primary motor cortex controlling left upper limb muscles (first dorsal interossei) (Magnani et al., 2014).

Somatosensory electrical stimulation is a promising and accessible intervention to enhance ascending afferent inputs and improve spatial representation from the specific body region where stimulation is applied (Guariglia et al., 1998; Kerkhoff, 2003; Pitzalis et al., 2013; Vallar et al., 1995). Several studies show that sensory stimulation delivered to a dermatome can increase the excitability of M1 of the corresponding myotome (Charlton et al., 2003; Hamdy et al., 1998; Kaelin-Lang et al., 2002; Ridding et al., 2000; Ridding et al., 2001; Sorinola et al., 2012; Veldman et al., 2014). Previous reviews have advocated for the use of sensorimotor stimulation to rehabilitate spatial neglect deficits (Kerkhoff, 2003). Transcutaneous sensory stimulation of the left hand (Vallar, 1997; Vallar et al., 1995) and vibratory stimulation of the left neck (Guariglia et al., 1998) have been shown to improve neglect deficits.

Beyond peripheral stimulation, previous research demonstrates that transcranial direct current stimulation (tDCS) applied over the primary motor cortex (M1) can modulate the magnitude and duration of PA effects, or even reactivate previously learned PA aftereffects both in neurologically unimpaired adults (Panico et al., 2021; Panico et al., 2017) and patients with spatial neglect (O'Shea et al., 2017). These studies showed enhanced consolidation and longer-lasting retention of PA-induced behavioral improvements when anodal tDCS to M1 was applied anodal tDCS to M1 during or after prism exposure (O'Shea et al., 2017; Panico et al., 2017). In complementary work, Yoon et al. applied repetitive transcranial magnetic stimulation (TMS) to the primary somatosensory cortex during PA in healthy adults to create a transient, artificial somatosensory suppression, and showed that somatosensory gating contributes to the realignment processes engaged by PA (Yoon et al., 2014). Collectively, prior findings that noninvasive stimulation of somatosensory cortex can influence PA-related processing and also support a causal contribution of M1 excitability to the expression of prism aftereffects (Panico et al., 2017). Together these findings suggest a bidirectional relationship whereby PA training modifies M1 function and, conversely, enhancing M1 activity can amplify or prolong the sensorimotor and cognitive benefits of PA.

Overall, peripheral somatosensory stimulation and cortical neuromodulation converge on a common idea: augmenting somatosensory and motor circuit excitability can shape the magnitude and persistence of PA aftereffects. Here, we posit that sensory stimulation can be a low-cost, non-invasive, clinically-feasible treatment strategy to enhance contralateral M1 excitability and augment the neurophysiologic effects of PA. However, no previous studies have concurrently investigated the neurophysiological and behavioral effects of PA combined with stimulation in the same participant cohort. From a theoretical perspective, combining PA with somatosensory stimulation offers a direct test of whether somatosensory upregulation can amplify PA-induced plasticity within the frontoparietal and sensorimotor networks. Clinically, combining PA with somatosensory stimulation could strengthen and potentially help transfer PA aftereffects to non-trained muscles, which can enhance current rehabilitation protocols to provide larger system-level gains rather than limb-specific improvements. Finally, quantifying both behavioral and TMS-based neurophysiological effects allows a deeper mechanistic evaluation of the added value of PA combined with stimulation beyond PA alone. To address these gaps, in the current study, we evaluate the neurophysiological and behavioral effects of noninvasive transcutaneous somatosensory nerve stimulation as a feasible, noninvasive, accessible, and clinically translatable adjunct to PA.

The inter-effector transfer of behavioral and neurophysiological effects of PA also merits more in-depth investigation. The leftward aftereffect observed after exposure to right-shifting prisms is thought to be mediated by changes in cortical networks responsible for spatial and attentional processing, which have connectivity with the primary motor cortex (Bartolomeo et al., 2012). Notably, well-established PA protocols for treatment of spatial neglect combine right-shifting prism exposure with unilateral upper limb motor training (Luauté, Halligan, et al., 2006; Rossetti et al., 1998). However, PA likely leads to a global modulation of visual and spatial cortical networks, as well as their corresponding connections to sensory and motor cortical areas, and thus, could induce a recalibration of both upper and lower limb sensorimotor circuits. Conflicting results regarding the inter-effector transfer of PA aftereffects between limbs (e.g. from the upper to the lower limbs) underscore key unresolved questions (Morton & Bastian, 2004; Savin & Morton, 2008). Additionally, the scarcity of prior PA studies that incorporate both neurophysiological and behavioral outcomes, highlight important unanswered questions about the aftereffects of PA, and whether PA with upper limb motor training would elicit measurable aftereffects in the lower limb. Addressing these gaps could provide critical insights for optimizing limb positioning and the choice of appropriate motor training tasks during PA.

Thus, the objective of the current study was twofold. First, we tested the hypothesis that following a single session of right-shifting PA, behavioral and neurophysiological effects would be enhanced by the addition of somatosensory stimulation to the left upper limb during prism exposure. Second, we hypothesized that right-shifting PA would produce inter-effector transfer of sensorimotor aftereffects to both the left upper and lower limbs (toward the direction of the leftward prismatic after-effect). We wanted to evaluate whether PA results in global effects, inducing measurable effects on both upper and lower limb excitability or local effects restricted to the limb engaged in the motor task during PA, i.e. right upper limb. Here, as an important first step, we tested these hypotheses in a group of young able-bodied individuals to elucidate the neural and behavioral mechanisms of PA combined with somatosensory simulation (Stim) or Sham in the neurologically unimpaired nervous system.

2. Methods

Our study design comprised a single-blind, repeated-measures, 2-arm randomized, crossover comparison of the behavioral and neurophysiological effects of two PA conditions: PA combined with sham stimulation (PA+Sham) and PA combined with somatosensory electrical stimulation (PA+Stim). The experimental protocol was approved by the Emory University Institutional Review Board and registered on clinicaltrials.gov. All participants provided written informed consent for study participation.

2.1. Participants

Young able-bodied participants were recruited for the study from the university and local community (Table 1). Participants were excluded if they had a history of or evidence of orthopedic or neurological pathology, pregnancy (female), presence of skin conditions, bruises, or cuts on the left upper limb at the electrical stimulation electrode placement site, impaired sensation in the left upper limb, concurrent enrollment in an upper limb rehabilitation study, any medical conditions interfering with study procedures, cardiac pacemaker or other implanted electronic devices, and any contraindications to TMS (Rossi et al., 2021). All participants had normal or corrected-to-normal vision. Sixteen young able-bodied individuals (9 females, 7 males, age 26.5± 4.4 years) completed the study (Table 1). One participant withdrew from the study due to discomfort during TMS, providing complete study data on N=15 (8 females, 7 males, age 26.7±4.4 years). The average time between the 2 PA sessions was 27±8.7 days (maximum duration between the 2 sessions was 41 days). Handedness was determined based on self-report, and 1 participant was left-handed.

Table 1:

Study participant demographics

	Age (Years)	Gender	Race	Order of Stim		Time between sessions (Days)	Handedness
				1st session	2nd session
1	35	Male	White	Sham	Stim	28	Right
2	21	Female	Asian	Stim	Sham	16	Right
3	22	Female	White	Stim	Sham	20	Right
4	25	Female	White	Sham	Stim	20	Right
5	26	Female	Black	Sham	Stim	21	Right
6	30	Male	Black	Sham	Stim	21	Left
7	27	Female	White	Stim	Sham	29	Right
8	30	Female	Black	Sham	Stim	18	Right
9	27	Male	Asian	Stim	Sham	45	Right
10	28	Male	Asian	Sham	Stim	26	Right
11	21	Male	Asian	Stim	Sham	35	Right
12	35	Male	Asian	Sham	Stim	28	Right
13	23	Female	White	Stim	Sham	33	Right
14	26	Male	Black	Sham	Stim	19	Right
15	24	Female	White	Stim	Sham	41	Right
	26.7± 4.4	7M, 8F	-	-	-	27±8.7	-

2.2. Study Procedures

Participants completed two experimental sessions comprising PA+Sham and PA+Stim, with >3-week wash-out period between sessions, due to previous data showing the effects of 1-2 sessions can last up to 1 week, to ensure that there are no carryover effects from session 1 to session 2 (Farnè et al., 2002; Fortis et al., 2011; Luauté, Michel, et al., 2006; Rossetti et al., 1998). We randomized the order of sessions such that half the participants completed PA+Sham first and the other half completed PA+Stim first (Figure 1). Participants were blinded to whether they received Stim or Sham during PA. All participants were given the same instructions during both sessions and were told that they might or might not feel the stimulation. Each experimental session comprised 3 parts in the following order:

Figure 1:

A) Overview of study design: Participants were randomized to receive one session of PA combined with either Stim (PA+Stim) or Sham (PA+Sham) in a randomized order with at least a 3-week washout period between the two conditions. Behavioral (visuo-proprioceptive and proprioceptive pointing) and neurophysiological (single and paired-pulse TMS-evoked responses) were obtained immediately Pre and Post each session of PA. B) Representative behavioral data during the upper extremity pointing task with eyes open before (without prism glasses on), during (with rightward shifting prism glasses on), and after PA (without prism glasses on). C) Representative neurophysiological data showing TMS-evoked MEP responses from the left FDI muscle at pre and post time points in the PA+Sham (left) and PA+Stim (right) conditions.

1. Pre-PA measurement of corticomotor neurophysiology using TMS, followed by measurement of visuospatial-motor behavior using two upper limb pointing tasks, one with eyes open and the other with eyes closed.

2. PA session with somatosensory or sham stimulation (PA+Stim or PA+Sham)

3. Post-PA measurement of upper limb pointing tasks and corticomotor neurophysiology, using the same methods as the Pre-PA measurement.

2.2.1. Visuospatial-motor Behavior Measurements:

To assess the effects of PA on visuospatial-motor behavior during upper limb pointing tasks, the Kessler Foundation Prism Adaptation Treatment (KF-PAT^®) protocol and devices (Stoelting Co., IL, USA) were used (Chen et al., 2015).

2.2.2. Prism Adaptation and Pointing Tasks during PA

Prism exposure: During prism exposure, participants wore a 20-diopter left-base wedged prism lenses that induced a rightward 12.4° visual field (Bernell^™ Deluxe Prism Training Glasses) fitted in rubber goggles that blocked peripheral vision views. A visual field occluder blocked the participant’s view of their proximal arm movements. During PA, participants sat comfortably with their right upper limb positioned for pointing movements towards visual targets presented in front of them at a distance of about 55cm from the trunk.

Pre and post PA aftereffect assessment:

We measured 10 proprioceptive pointing (eyes closed) and 5 visuoproprioceptive pointing trials (eyes open) immediately before and after PA. The proprioceptive pointing task required the participants to point straight from their chest with the right upper limb until full arm extension while keeping their eyes closed. The visuo-proprioceptive pointing task required participants to point straight from their chest to visual targets placed at their arm’s length distance toward their left, right, and center hemi-space with their eyes open, but their arms and fingers were covered using a black cloth to obscure visual feedback from the entire upper limb

Motor practice during prism exposure:

Participants performed 120 pointing trials with their right arm fully extended toward visual targets located at 0⁰ or 21⁰ to the right or left of a board that was oriented to the participant’s median plane. Our current protocol replicates standard upper limb pointing practice by the right arm (non-paretic arm) employed during the rehabilitation of left spatial neglect using PA (Barrett et al., 2012). Following the framework of Prablanc et al(Prablanc et al., 2020), we define inter-effector transfer as PA-induced changes in behavior and corticomotor excitability in the untrained (i.e. the limbs that did not engage in motor practice during PA) left upper limb first dorsal interossei (FDI) and left lower limb (tibialis anterior (TA), soleus) muscles.

2.2.3. Participant Setup and EMG Placement

Surface electromyography (EMG) signals were recorded from bilateral first dorsal interossei (FDI) muscles (Kendall^© 24mm Ag/AgCl, H124sG) as well as left tibialis anterior (TA) and left soleus muscles (Biopac^© 35 mm A/AgCl, EL502). EMG sensors were attached to the skin overlying the muscle belly of each of the 4 muscles. EMG placement and signal quality were checked by asking the participants to activate each muscle while the experimenter monitored EMG activity. EMG data were collected at the sampling rate of 2000 Hz and band-pass filtered at 10-1000Hz using a 6-channel wireless EMG System (Biopac^© Systems Inc. MP160WSW, AcqKnowledge software).

2.2.4. Single and Paired Pulse TMS Measurements

During TMS measurements for the FDI muscles, participants were in a quiet seated position in a standard chair, back and thighs supported, knees bent at 90 degrees, feet flat on the floor, with both forearms and hands resting on a firm foam bar. TMS targeting FDI was delivered using a hand-held 70mm diameter figure of eight coil positioned at a 45⁰ angle from the midsagittal plane to target the FDI muscle M1 hotspot. For TMS measurements from the left TA and left soleus muscles, participants were in a quiet standing position, with both hands resting on a front support bar. TMS targeting leg muscles was delivered using a 70mm figure-of-eight custom batwing coil positioned to target the M1 hotspot of the left soleus muscle, with evoked responses recorded from both the left soleus and left TA. Single and paired TMS pulses were delivered using a Magstim 200² stimulator (MagStim Inc., Wales, UK). A real-time neuronavigation system which included the BrainSight MNI template (BrainSight, Rogue Research) was used to track and consistently maintain TMS coil position at the hotspot location throughout the experiment. The TMS hotspot was identified as the site on the contralateral motor cortex that generated the largest and most consistent motor evoked potentials (MEPs) from the respective targeted muscles.

For each participant, the M1 hotspot corresponding to each muscle was determined at the start of the first session using established TMS protocols (Groppa et al., 2012; Rossi et al., 2009; Rossini et al., 2015; Rossini & Rossi, 2007) and confirmed at subsequent sessions using the neuronavigation software. Based on the TMS hotspot identified, the resting motor threshold for bilateral FDI and active motor threshold for the left soleus were determined using the maximum likelihood model of parameter estimation by sequential testing (PEST) software (Awiszus, 2003). Resting motor threshold was determined as the minimum TMS intensity that generated ≥0.05mA MEP peak-to-peak amplitude for bilateral FDI, and active motor threshold for TA and soleus was determined as the minimum intensity that generated ≥0.1mA MEP amplitude for the soleus (Awiszus, 2003). A total of 10-20 suprathreshold TMS pulses (130% of active or resting motor threshold) were delivered to measure corticomotor excitability. The average number of pulses delivered was 15 per hotspot, and was determined by the tolerance of participants to the TMS. Next, three sets of paired pulse TMS trials were collected to obtain measures of intracortical circuit excitability- intracortical facilitation (ICF), short-interval intracortical inhibition (SICI), and long-interval intracortical inhibition (LICI). The paired-pulse stimulation parameters for SICI and ICF comprised conditioning TMS pulse intensity of 80% resting motor threshold (FDI) or 70% of active motor threshold (soleus and TA), and test TMS pulse intensity of 130% motor threshold(Chen, 2000; Leung et al., 2018). The inter-stimulus interval was 12ms for ICF and 2ms for SICI. The paired pulse TMS parameters for LICI were conditioning pulse intensity 130% motor threshold, test pulse intensity 130% motor threshold, and interstimulus interval of 100ms (Hand et al., 2020; Leung et al., 2018; Temesi et al., 2017).

Post-PA TMS data were collected using the same hotspot and TMS intensities as Pre-PA TMS measurements. During the second session, we re-registered landmarks in BrainSight for accuracy and we used the same TMS hotspot from the initial session using neuronavigation, confirmed the hotspot location, and the MEP thresholds were determined for each session.

2.2.5. Electrical Stimulation and Sham Stimulation

Surface rectangular electrical stimulation electrode pads (2” x 4”) were placed on the left upper limb over the belly of the triceps and biceps muscles in the left upper arm, and smaller 2” x 2” square pads were placed on the ventral aspect of the left forearm muscles. Electrical stimulation intensity was customized according to each participant’s sensory threshold, determined as the lowest intensity at which the stimulation was perceived, and motor threshold, determined as the lowest intensity that elicited a visible muscle twitch(Sorinola et al., 2012; Veldman et al., 2014). During PA+Stim, stimulation was determined in a customized manner for each participant at approximately the midpoint of sensory and motor threshold intensities (i.e. above sensory and below motor threshold) for the duration of 20 minutes during prism exposure (Sorinola et al., 2012; Veldman et al., 2014). During PA+Sham, stimulation electrodes were placed on the arm and forearm as described above, stimulation threshold was determined, and then the stimulation device was turned off. Stimulation was delivered using an Empi TENS unit using standard device parameters for TENS, at a pulse rate of 100 Hz, and pulse width of 300μsec.

2.3. Data Analyses

Primary behavioral dependent variables included the magnitude of deviation with respect to the midline during visuo-proprioceptive pointing and proprioceptive pointing tasks performed with the right upper limb, with deviation toward the left of midline measured as negative values (distance in cm) and to the right as positive values. Primary neurophysiological dependent variables included suprathreshold MEP amplitude for each of the 4 muscles (left FDI, right FDI, left soleus, left TA). Secondary variables included PA-induced change in SICI, LICI, and ICF (calculated as the percentage of the conditioned MEP amplitude versus the unconditioned MEP amplitude) for each of the 4 muscles. Each participant’s neurophysiology data were analyzed to determine the peak-to-peak MEP amplitude using a custom TMS processing software(Cunningham et al., 2021). Sample size estimation was conducted a priori using preliminary data for the primary variable (left FDI MEP amplitude) in G-power, which showed effect sizes of Cohen’s d ≈ 0.69 at α=0.05. For each of the 6 primary variables, we conducted repeated-measures 2-way ANOVAs to evaluate the effects of training condition (Stim PA, Sham PA) and time (Pre, Post) on each dependent variable. Also, for all variables (primary and secondary), we conducted a priori planned paired t-tests to compare the pre to post magnitude of change or percentage (%) change (Post minus Pre)/Pre) x 100) between the PA+Stim and PA+Sham conditions. We used thresholds of p<0.05 for significance. We used IBM SPSS^© statistics software for statistical analyses (Version 29.0.2.0 (20)).

3. Results

3.1. Primary Behavioral Outcome Measures: Proprioceptive and Visuo-proprioceptive Pointing

For proprioceptive pointing, the 2-way repeated measures ANOVA showed that there was a significant main effect of time (F (1,14) = 30.66, ηp² = 0.69, p = < 0.001), no significant main effect of condition (F (1,14) = 1.53, ηp² = 0.1, p= 0.24), and no significant condition by time interaction (F (1,14) = 0.22, ηp² = 0.16, p= 0.65) (Figure 2A). The a priori planned paired t-test comparing pre-post change scores between the PA+Stim and PA+Sham conditions showed no significant difference in the magnitude of Pre to Post change for PA+Stim (−4.00 ± 13.27 SD) versus the PA+Sham (−3.45 ± 13.49 SD) condition (t= −0.47, ηp² =0.02, p= 0.65) (Figure 2B).

Figure 2: Effects of PA combined with Stim and Sham on visuospatial-motor behavior:

A) Visuo-proprioceptive pointing (eyes open) performance during the PA+Sham and PA+Stim conditions at pre and post timepoints. Note that there was a significant leftward aftereffect at post versus pre, as shown by a greater negative shift from pre to post for both the PA+Stim condition (p= <0.001) and the PA+Sham condition (p= <0.001). B) Proprioceptive pointing (eyes closed) performance during PA+Sham and PA+Stim conditions at the pre and post timepoints. We observed a significant leftward aftereffect, as shown by a shift or deviation toward the left space from pre to post for both the PA+Stim condition (p= <0.001) and for the PA+Sham condition (p= <0.002). C) Graphs comparing change scores (post minus pre) for visuo-proprioceptive (eyes open) and proprioceptive pointing (eyes closed) for the PA+Sham and PA+Stim conditions. There was no significant difference in magnitude of PA-induced change between the 2 conditions.

For visuo-proprioceptive pointing, there was a significant main effect of time (F (1,14) = 100.84, ηp² = 0.88, p <0.001), no significant effect of condition (F (1,14) = 0.43, ηp² = 0.03, p = 0.52), and no significant condition by time interaction (F (1,14) = 0.03, ηp²= 0.00, p = 0.88). The a priori planned paired t-test comparing pre-post change scores between the PA+Stim and PA+Sham conditions showed no significant difference in the magnitude of training-induced change for PA+Stim (−3.52 ± 6.55 SD) versus PA+Sham (−3.49 ± 7.37 SD) condition (t=−0.06, ηp² =0.00, p= 0.96,) (Figure 2B).

3.2. Primary Neurophysiological Outcome Measures: TMS-evoked MEPs of bilateral upper limb and left lower limb muscles

For MEP amplitude of the left FDI, the 2-way ANOVA showed a significant effect of time (F (1,13) = 5.33, ηp² = 0.29, p= 0.04), no significant main effect of condition (F (1,13) = 0.39, ηp² = 0.03, p = 0.54), and a significant condition by time interaction (F (1,13) = 6.33,ηp² = 0.33, p = 0.03). The post-hoc paired comparisons showed that left FDI MEP amplitude significantly increased from pre to post for the PA+Stim condition (ηp² = 0.52, p=0.002) but showed no significant change from pre to post for the PA+Sham condition (ηp² = 0.00, p= 0.94). The a priori planned paired t-test comparing Pre to Post change scores in MEP amplitude between the PA+Stim and PA+Sham conditions showed a significantly larger magnitude of change in left FDI MEP amplitudes for the PA+Stim (+0.74mV ± 0.79 SD) vs the PA+Sham (+0.18mV ± 0.56 SD) condition (t (1,14) = 2.48, ηp²= 6.88, p = 0.03) (Figure 3).

Figure 3: Effects of PA combined with Stim and Sham on corticomotor excitability of bilateral upper extremity (FDI) muscles.

A) MEP amplitudes at Pre and Post timepoints for the PA+Sham and PA+Stim for the left FDI and right FDI muscles. Left FDI MEPs showed a significant increase at Post versus Pre in the PA+Stim condition (p= 0.002). B) Training-induced change (post minus pre) in left FDI and right FDI MEP amplitudes for the PA+Sham and PA+Stim conditions. We found a significantly larger magnitude of change in left FDI MEP amplitudes for the PA+Stim vs PA+Sham condition (p= 0.029, t= 2.482). There was no significant difference between PA+Sham and PA+Stim in the magnitude of change in right FDI MEP amplitudes.

For MEP amplitudes of the right FDI, there was no significant main effect of time (F (1,14) = 4.46, ηp²= 0.24, p= 0.05), no significant effect of condition (F (1,14) = 0.67, ηp²= 0.05, p= 0.43), and no significant condition by time interaction (F (1,14) = 0.63, ηp²= 0.04, p= 0.44). The a priori planned paired t-test comparing Pre to Post change scores between the PA+Stim and PA+Sham conditions showed no significant difference in the magnitude of change for the PA+Stim (0.31mV ± 0.85 SD) versus PA+Sham (0.41mV ± 0.85 SD) condition (t = −0.79, ηp² = −0.08, p = 0.44) (Figure 3).

For left TA MEP amplitudes, there was no significant main effect of time (F (1,12) =3.16, ηp²= 0.21, p= 0.10), no significant main effect of condition (F (1,12) = 2.52, ηp²= 0.17, p = 0.14), and no significant condition by time interaction (F (1,12) = 0.12, ηp²= 0.01, p= 0.74). The a priori planned paired t-test comparing Pre to Post change scores showed no significant difference in the magnitude of Pre to Post change in TA MEP amplitudes for the PA+Stim (0.11mV ± 0.16 SD) versus PA+Sham (0.11mV ± 0.46 SD) condition (t (1,11) = −0.03, ηp²= 0.00, p = 0.98) (Figure 4).

Figure 4: Effects of PA combined with Stim and Sham on corticomotor excitability of left lower extremity (soleus, TA) muscles.

A) MEP amplitudes at Pre and Post timepoints for the PA+Sham and PA+Stim for the left soleus and left TA muscles. There was no significant difference in left soleus and TA MEPs between the 2 conditions. B) Training-induced change (post minus pre) in left soleus and TA MEP amplitudes for the PA+Sham and PA+Stim conditions.

For left soleus MEP amplitudes, there was no significant main effect of time (F (1,12) = 0.55, ηp²= 0.04, p= 0.47,), no significant main effect of condition (F (1,12) = 2.17, ηp²= 0.15, p = 0.17, nor condition by time interaction (F (1,12) = 3.32, ηp²= 0.22, p= 0.09). The a priori planned paired t-test comparing Pre to Post change scores in MEP amplitude between the PA+Stim and PA+Sham conditions showed no differences in the magnitude of change in left soleus MEP amplitudes for the PA+Stim (+0.07mV ± 0.08 SD) versus the PA+Sham (−0.02mV ± 0.18 SD) condition (t (1,12) = 1.82, ηp²= 0.5, p = 0.09) (Figure 4).

3.3. Secondary Neurophysiologic Variables: SICI and LICI

For Pre to Post % change in SICI of the left TA, the paired t-test showed a significant difference in the magnitude of change for the PA+Sham (+21.36% ± 36.92 SD) versus PA+Stim (−29.18% ± 26.01 SD) condition (t = −3.24, ηp²= −1.03, p = 0.01). For the left FDI, right FDI, and left soleus muscles, the paired t-test comparing Pre to Post change in SICI showed no significant differences between the PA+Stim and PA+Sham conditions (all p-values >0.1) (Figure 6).

For Pre to Post change in LICI of the left FDI, the paired t-test comparing Pre to Post % change scores showed a significant difference in the magnitude of change for the PA+Sham (+91.29% ± 179.62 SD) versus PA+Stim (−26.87% ± 56.47 SD) condition (t = −2.26, ηp² = −0.65, p = 0.045). For change in LICI of the left TA, the paired t-test comparing % change scores showed no significant difference between the PA+Sham (+26.28%) versus PA+Stim (−24.16) condition (t = −1.74, ηp²= −0.55, p = 0.12). For the right FDI and left soleus muscles, the paired t-test comparing Pre to Post % change in LICI showed no significant difference between the PA+Stim and PA+Sham conditions (all p-values >0.1) (Figure 7). Results and corresponding figures for ICF are presented in the supplementary materials.

4. Discussion

We evaluated the behavioral and neurophysiological effects of a single session of PA with right-shifting prism glasses, combined with either somatosensory electrical stimulation (PA+Stim) or sham stimulation (PA+Sham) delivered to the left upper limb (targeting the direction of the PA aftereffect); and we investigated how this combination shaped corticomotor excitability, intracortical inhibition, and inter-effector transfer in young healthy adults. To our knowledge, this is the first study evaluating both behavior and corticomotor neurophysiology in the context of PA in both upper and lower limb muscles.

Our results showed that in young able-bodied adults, a single session of both PA+Stim and PA+Sham induced a significant visuospatial-motor behavioral aftereffect, evidenced by a significant post-PA leftward shift during visuo-proprioceptive and proprioceptive pointing tasks. Notably, there were no significant differences in the magnitude of these behavioral aftereffects between the PA+Stim versus PA+Sham conditions. In contrast, only PA paired with stimulation increased corticomotor excitability in the left upper limb, with significantly larger PA-induced change in FDI MEP amplitude when PA was combined with Stim versus Sham, and PA-induced changes in left FDI LICI and left TA SICI also differed between PA+Stim and PA+Sham. Below, we first interpret these neurophysiologic and behavioral results in relation to current mechanistic models of PA, and then discuss their potential clinical implications and relevance for rehabilitation.

Our neurophysiological measures revealed that only PA paired with stimulation resulted in increased corticomotor excitability in the left upper limb, with significantly larger PA-induced change in FDI MEP amplitude when PA was combined with Stim versus Sham. Furthermore, PA-induced changes in Left FDI LICI and Left TA SICI showed significant differences between PA+Stim versus PA+Sham, demonstrating differential effects on intracortical inhibition in the left upper and lower limbs. Our observation that PA+Stim altered corticomotor excitability in both the trained (right upper limb) and untrained muscles (e.g. change in left TA SICI) aligns with the bidirectional M1-PA relationship established by recent neurostimulation work, where M1 may serve as an anatomo-functional interface within a hierarchical adaptation network that can store latent sensorimotor adaptations (O'Shea et al., 2017; Panico et al., 2017). Our results suggest that peripheral somatosensory stimulation may enhance PA-induced plasticity (increased MEP amplitude and LICI in left FDI, increased SICI in Left TA) by similarly engaging this bidirectional coupling, selectively amplifying neurophysiological outcomes despite the lack of differences in behavioral aftereffects in a single session.

Our results showed that right-shifting PA, whether combined with Stim or Sham, induced a robust leftward behavioral aftereffect during upper limb pointing movements, resulting in a leftward shift in visuospatial orientation during upper limb pointing tasks in young healthy adults. Our current results are consistent with well-established behavioral aftereffects of PA reported in previous literature (Redding et al., 2005; Rossetti et al., 1998) and further support the validity of PA as an intervention targeting visuospatial orientation and sensorimotor function (Chen P 2019; Magnani et al., 2014; Rossetti et al., 1998; Tilikete et al., 2001; Watanabe & Amimoto, 2010). Contrary to our hypothesis that augmenting somatosensory afferent activity from the left upper limb (toward the side of the prism aftereffect) would enhance the behavioral aftereffects of PA, we observed no significant difference in the magnitude of PA-induced leftward shift between PA+Stim versus PA+Sham. Thus, our results suggest that adding somatosensory stimulation to PA may not induce marked augmentation of PA’s behavioral effects within only one session. Presumably, higher dosage, such as multiple sessions of PA combined with Stim, may be required to enable the synergistic enhancement of the behavioral effects of PA. Furthermore, future work can potentially evaluate subtle behavioral differences with tools such as marker-based or markerless video-based motion capture.

In our current study on able-bodied healthy adults, we hypothesized that following PA, there would be an increase in corticomotor excitability in the hemisphere contralateral to the leftward prism aftereffect (i.e. increased left FDI MEP amplitude). Our results partially supported our hypothesis. PA induced an increase in corticomotor excitability of the left FDI, but only following PA+Stim and not after PA+Sham. Furthermore, the training-induced change in left FDI MEP amplitude was significantly larger with PA+Stim compared to PA+Sham. Our findings regarding the absence of increased corticomotor excitability with PA+Sham are consistent with previous studies showing that rightward shifting PA (leftward aftereffect) did not increase corticomotor excitability in the left FDI in young able-bodied adults (Magnani et al., 2014; Schintu et al., 2016). Another previous study demonstrated that a single session of left M1 anodal (excitatory) transcranial direct current stimulation (tDCS) applied during right shifting PA enhanced PA neurophysiological aftereffects, whereas PA without stimulation did not induce changes in corticomotor excitability (O'Shea et al., 2017). Also, our findings of lack of changes in the right FDI suggest that PA did not increase interhemispheric asymmetry by reducing corticomotor excitability of the right hemisphere. We found no significant change in right FDI corticomotor excitability following either PA+Sham or PA+Stim, which aligns with a prior study showing no change in ipsilateral corticomotor excitability following right-shifting PA (Magnani et al., 2014). Also, Arévalo et al. reported interesting asymmetries in young able-bodied individuals between the effects of leftward and rightward shifting PA on inter-hemispheric inhibition (Martín-Arévalo et al., 2018), which somewhat align with the lack of changes in right FDI in our study and merit follow-up in larger sample studies.

In multiple previous studies, sensorimotor electrical stimulation has been shown to increase afferent input in the somatosensory cortex (Rosenkranz & Rothwell, 2003, 2004), which in turn enhances the excitability of corresponding motor cortical areas (M1) (Charlton et al., 2003; Kaelin-Lang et al., 2002; Ridding et al., 2001). Based on these findings, we predicted that PA+Stim (with Stim delivered to the left upper limb) would enhance the aftereffects of PA. Indeed, our results showed that PA+Stim increased left FDI MEP amplitudes, whereas PA+Sham did not. Furthermore, compared to PA+Sham, PA+Stim produced a significantly larger training-induced change in MEP amplitude of the left FDI (Figure 1B). Based on our findings, we suggest that combining PA with sensory electrical stimulation targeting the upper limb on the side of the prismatic aftereffect (left upper limb) may strengthen inputs from the contralateral right somatosensory to the right motor cortical areas, resulting in greater MEP amplitudes in left upper limb muscles. However, despite these larger neurophysiological effects of PA+Stim on left FDI cortical excitability, we did not observe correspondingly greater behavioral aftereffects following PA+Stim compared to PA+Sham. Taken together, these mechanistic results suggest that somatosensory stimulation might alter how PA-related recalibration is implemented within M1 microcircuits, rather than changing the magnitude of the immediate behavioral aftereffect. Potentially, the increased corticomotor output induced by PA+Stim provide novel insights regarding subtle neurophysiologic effects that could contribute to measurable behavioral changes if further amplified. Greater dosage of PA (e.g. 5-10 sessions) or stimulation (e.g. stimulation to both upper limb and lower limb) may elicit larger behavioral aftereffects with PA+Stim. The additive neural and behavioral effects of PA combined with somatosensory Stim merit future research.

Interestingly, our results provide some support for the prediction that neurophysiological aftereffects of PA+Stim combined with upper limb motor training may transfer to the left lower limb. Based on the hypothesized mechanisms through which PA modulates cortical neural networks, we predicted that rightward shifting PA would increase excitability within left visuospatial and attentional networks, thereby facilitating projections from the parietal cortex onto sensory and motor cortical areas more globally. Thus, we anticipated that the effects of PA would show inter-limb transfer to the left lower limb, and not just be localized to the upper limbs, when PA is combined with upper limb motor training. While a few previous studies evaluated the behavioral inter-effector transfer of PA between upper and lower limbs or during gait (Savin & Morton, 2008), to our knowledge, the neurophysiologic inter-effector transfer of PA effects are poorly understood.

Several studies have demonstrated that somatosensory stimulation to a dermatome can increase corticomotor excitability of the homologous myotome (Hamdy et al., 1998; Kaelin-Lang et al., 2002; Ridding et al., 2001; Sorinola et al., 2012; Veldman et al., 2014). For example, Mang et. al. reported that high-intensity motor-level neuromuscular electrical stimulation to upper limb muscles had a focused effect on the target or stimulated muscle MEP amplitudes alone and did not transfer to leg muscles (Mang et al., 2011). Similarly, Kato et. al. found that neuromuscular electrical stimulation delivered to wrist muscles (upper extremity Stim) did not cause inter-effector transfer of upregulated corticomotor excitability to leg (Kato et al., 2022). Taken together, these previous reports suggest that somatosensory stimulation delivered to the left upper limb alone should not increase corticomotor excitability in the left leg muscles. Thus, in our study, where we combined sensory stimulation delivered to the left upper limb with right-shifting PA, we posited that the combination of PA+Stim will facilitate the inter-effector transfer of the effects of Stim to non-stimulated and non-trained lower limb muscles. Such a transfer effect of PA+Stim could be advantageous for enhancing the therapeutic impact of PA protocols. Our results partially align with our hypothesis, suggesting that when somatosensory stimulation is added to PA, the stimulation-induced changes in neural circuit excitability may augment PA-induced neuroplasticity, increasing the likelihood of transfer to lower limb muscles, which is supported by our results related to differential PA-induced changes in Left TA SICI with versus without Stim. To our knowledge, no other studies have evaluated the effect of PA+Stim on neurophysiological outcomes of both upper and lower limb muscles.

To confirm that PA combined with somatosensory stimulation can promote a greater magnitude of PA effects or greater inter-effector transfer from upper to lower limb corticomotor output, future studies could explore several interesting research directions. For example, we could investigate the effects of combining PA with a lower limb motor training task or targeted somatosensory stimulation to the left lower limb, which could further augment carryover effects to lower limb muscles, and potentially enhance the behavioral aftereffects of PA. Our results also underscore the need for further research on lower limb neurophysiology following PA in neurologically unimpaired individuals and those with spatial neglect post-stroke. Moreover, because behavioral changes may accompany corticomotor excitability, future investigations should also examine the effects of PA on lower limb motor performance tasks such as seated leg pointing tasks or foot placement during walking.

We found that compared to PA+Sham, PA+Stim induced a significantly larger increase in LICI of the left FDI, as well as a significantly larger increase in SICI of the left TA, indicating greater intra-cortical inhibitory influences on M1 after PA+Stim. These findings suggest that combining PA with somatosensory stimulation may facilitate GABAA interneurons within the motor cortex, increasing intracortical inhibition. Other studies have previously shown that somatosensory stimulation alone induces GABA-mediated changes in SICI (Celnik et al., 2007; Conforto et al., 2010; Farias da Guarda & Conforto, 2014). Bracco et al. (Bracco et al., 2018)demonstrated that the combination of rightward shifting PA and excitatory tDCS of the right M1 led to reduced corticomotor excitability of the left upper limb. Our current neurophysiology results may be in congruence with and help expand the literature that combines PA with other types of stimulation or neuromodulation. For example, studies incorporating neck vibratory stimulation (Choi & Lee, 2022), as well as left (Vallar, 1997) and right (Làdavas et al., 2015) M1 tDCS have reported enhanced effects of PA on spatial neglect deficits and motor function in post-stroke individuals with spatial neglect (He et al., 2007). Clinically, the robust behavioral and neurophysiological aftereffects of PA (both with and without Stim) in the left FDI validate its therapeutic potential for spatial neglect rehabilitation, where multiple sessions improve visuospatial attention and reduce contralateral deficits.

4.1. Limitations

There are a few limitations to the current study. Consistent with previous PA protocols, our behavioral training task during PA was performed exclusively with the right arm. This approach aligns with paradigms used in stroke participants with left spatial neglect, whose left arm movement may be limited due to paresis (Goedert et al., 2015; Rossetti et al., 1998). Future work can evaluate the effects of PA with the left versus right arm engaged in motor practice in able-bodied individuals. Although the participants were blinded to the stimulation condition, we did not perform a post-session check of blinding efficacy, so could not confirm whether participants had any insights into the stimulation condition (real vs sham). This limitation was partially mitigated by providing all participants identical instructions that they might or might not feel any stimulation and scheduling an adequate washout interval between both sessions to minimize carryover or expectation effects. Additionally, because the average experimental session duration ranged from three to four hours, fatigue could have affected the post-PA behavioral and neurophysiological data in some participants. We did not include a stimulation-only condition, preventing us from specifically isolating the effects of peripheral stimulation alone. This decision was made to minimize the number of experimental sessions and participant burden, due to the study being part of a registered clinical trial with a pre-defined protocol, and due to support from existing literature already demonstrating the robust effects of sensory stimulation on corticomotor excitability (Kato et al., 2022; Ridding et al., 2001; Sorinola et al., 2012; Veldman et al., 2014). However, the current methods do not allow us to parse out the effects of Stim alone from the combined effects of PA+Stim. Additionally, we did not consider handedness as a primary covariate in our primary analyses due to our modest sample size, only one left-handed person in our participant cohort, and thus the effects of handedness being beyond the scope of our study. Future studies may systematically examine handedness effects on PA inter-effector transfer, particularly given clinical relevance to heterogeneous stroke populations, where left neglect often co-occurs with variable hand preference. Finally, the current study evaluated young, neurologically unimpaired individuals, and future studies should explore the cognitive, behavioral, and neural aftereffects of PA in post-stroke individuals with spatial neglect.

4.2. Conclusions

Using a dose-matched, repeated-measures, sham-controlled study design, our findings suggest that while both PA+Stim and PA+Sham produce similar behavioral effects, the addition of somatosensory electrical stimulation may yield broader and more robust neurophysiological effects compared to PA alone. Our study demonstrated the feasibility of combining PA with Stim and provided new evidence that right-shifting PA, when combined with somatosensory electrical stimulation delivered to the left upper limb, increases corticomotor excitability of the left upper limb (toward the direction of the leftward prism aftereffect) in young healthy adults. Notably, our results also revealed that the neural effects of PA+Stim may transfer from the upper limb to the left lower limb in young healthy young adults (change in intracortical inhibition), highlighting a potential mechanism of interlimb transfer. These findings open avenues for future research to evaluate whether PA+Stim can enhance the therapeutic effects of PA in individuals with spatial neglect, particularly those individuals who do not respond to traditional PA alone. Furthermore, the potential for inter-effector transfer of neurophysiologic effects of PA to lower limb muscles warrants further investigation and could inform the development of targeted rehabilitative strategies to improve lower limb and walking function. Over the long-term, this line of research may guide the optimization of PA-based therapeutic protocols, incorporating neurobiological principles to enhance efficacy, and combining PA with adjuvant stimulation, neuromodulation, or motor training for clinical applications.

Figure 5: Effects of PA combined with Stim and Sham on short interval intracortical facilitation (SICI) and long interval cortical inhibition (LICI).

A) Graphs showing PA-induced % change (pre to post) in SICI for the PA+Sham and PA+Stim conditions for Left FDI, right FDI, left TA, and left soleus muscles. Only the left TA showed a significant difference in % change between PA+Sham versus PA+Stim condition (p= 0.01, t=−3.24). B) Graphs showing training-induced % change (pre to post) in LICI for the PA+Sham and PA+Stim in the left FDI, right FDI, left TA, and left soleus. Pre-post % change for left FDI LICI showed significant differences between the PA+ Sham versus PA+Stim condition (p= 0.045, t=−2.26).

Highlights.

• We evaluated behavioral and neural effects of prism adaptation (PA) combined with sensory stimulation.

• Both PA+Stim and PA+Sham induced significant behavioral aftereffects.

• PA+Stim induced larger increase in left upper limb corticomotor excitability and differential modulation of left lower limb intracortical inhibition.

• Effects of PA+ Stim may transfer from upper limb to lower limb