Influence of aging and task-related activation on descending cortical modulation of spinal sensorimotor circuitry

Abstract

The objective of this study was to evaluate the effects of aging on descending modulation of spinal reflexes across different task conditions. In healthy young (YA; age 27 ± 4 years) and older (OA; age 63 ± 10 years) adults, we utilized paired subthreshold transcranial magnetic stimulation (TMS) and peripheral nerve stimulation (PNS) to elicit unconditioned (PNS only) and conditioned (TMS paired with PNS) soleus Hoffmann (H-)reflexes at 3 inter-stimulus intervals (ISIs) (−1.5, +10 ms, +40 ms) and during 3 task conditions − sit rest (SR), sit active (SA), and quiet stance (QS). Our results showed that at the + 40 ms ISI, compared to YA, the OA group showed significant differences in modulation (and reversal from facilitation to inhibition) of soleus H-reflexes during both SA and QS tasks (p < 0.0001). Correlation analysis showed a significant negative relationship between age and the magnitude of H-reflex modulation at the + 40 ms ISI, such that older individuals showed a larger magnitude of H-reflex inhibition. In conclusion, altered task-related modulation of spinal reflexes at the + 40 ms ISI may reflect specific aging-related effects on spinal sensorimotor integration, likely mediated via complex interactions between the influence of polysynaptic slower-conducting descending pathways, sensory afferents, and local spinal circuits, on spinal reflex activity during motor task performance.

Introduction

Human movement, whether volitional or reflexive, requires the activation of spinal lower motor neurons (LMNs), the final common pathway in the motor system [[1], [2], [3], [4]]. However, spinal circuitry is complex, with spinal LMNs receiving multiple synaptic inputs from descending supraspinal tracts [[5], [6], [7], [8]], peripheral afferents [9,10], and local spinal circuits [[11], [12], [13], [14]]. The ability to comprehensively and non-invasively probe the influence of these synaptic inputs on spinal LMNs in humans remains challenging.

Transcranial magnetic stimulation (TMS) [[15], [16], [17]] and peripheral nerve stimulation (PNS) [[18], [19], [20]], when used in isolation, elicit motor evoked potentials (MEPs) and Hoffman’s reflexes (H-reflexes), which probe descending corticospinal and spinal sensorimotor circuits, respectively. However, to address knowledge gaps regarding interactions between descending cortical pathways and spinal segmental circuitry, previous studies have paired non-invasive cortical and peripheral nerve stimulation (TMS + PNS) [[21], [22], [23], [24], [25], [26]]. Common in these studies, a subthreshold TMS conditioning pulse is delivered at a timing interval before or after PNS, and the resulting modulation of soleus H-reflex amplitudes provides insights regarding the magnitude and direction of the influence of TMS-induced descending volleys on spinal reflex circuits [23,26,27]. TMS intensity [[28], [29], [30]], coil orientation [30], PNS intensity [23], and relative timing of paired stimuli [21,25,26] have been shown to affect the magnitude of TMS-conditioned spinal reflexes in healthy adults. While findings from previous studies pairing TMS and PNS have established a methodological framework for probing descending cortical influences on spinal sensorimotor circuitry, the effects of aging and task-related activation on these circuit interactions are poorly understood.

Aging can cause changes in corticospinal and spinal reflex circuitry. Older adults may show longer soleus H-reflex latencies [[31], [32], [33], [34]], decreased soleus Hmax/Mmax ratios [9,[34], [35], [36]], and longer MEP latencies [9,36,37], which may indicate aging-related declines in speed of signal transmission and synaptic efficacy. Several studies comparing TMS-evoked MEP amplitudes in young and older adults have yielded varied results [9,36,38,39]. However, some of these contrasting findings may be due to differences in task-related muscle activation, and warrant further investigation. Importantly, the effect of aging on the interactions between descending cortical motor pathways and spinal reflex circuits remains unclear.

While the effects of task-related activation (e.g., voluntary or postural activation) on corticospinal and spinal reflex excitability have been studied in young adults, few studies have investigated the effects of task-related activation on TMS-conditioning of spinal reflexes and in older adults. With changes in posture, young adults exhibit a decrease in H-reflex amplitude from the seated or prone posture to upright stance [35,40,41], which is assumed to reduce the risk of unexpected reflex activity that may impair balance control [42,43]. Several studies on subthreshold TMS-conditioned soleus H-reflexes in young adults evaluate subjects in a resting seated or prone position [28,29,44]. Other studies in young adults have investigated the magnitude of descending modulation of soleus H-reflexes during different motor tasks such as seated voluntary activation [22,23,27], upright standing [44,45], and antagonist muscle activation [21,24,46]. The modulation of spinal reflexes is also crucial for balance and gait control, allowing for rapid adjustments that are necessary to maintain postural stability and appropriate muscle activity across different phases of the gait cycle [47]. Here, we seek to understand the interplay between descending modulation of spinal reflexes and task-related activation, in the context of aging-related changes within cortical and spinal sensorimotor circuits.

The objectives of the current study were to investigate how interactions between descending cortical and spinal sensorimotor circuits are influenced by: 1) healthy aging, and 2) task-related activation. To accomplish these objectives, we compared TMS-conditioning of soleus H-reflexes in: 1) young versus older adults, and 2) among 3 different tasks – sitting at rest, sitting with voluntary activation, and quiet standing (postural activation). To evaluate behavioral significance of these sensorimotor circuit interactions, we also studied relationships between descending cortical modulation of spinal reflexes and dynamic balance ability, measured using a validated, clinically-feasible dynamic walking task [48].

Methods

Study participants

Twenty-one young (YA; 11 females; age 27 ± 4 years) and twenty-six older (OA; 14 females; age 63 ± 10 years) neurologically intact adults were recruited to complete the experimental protocol. Each participant provided written informed consent in accordance with the Declaration of Helsinki. Study procedures were approved by the Emory University Institutional Review Board (IRB#00083425).

Experimental design

Participants completed a single experimental session lasting approximately 3.5 h. The dominant leg was determined by asking participants their preferred leg when kicking a ball [49,50], and electromyographic (EMG) recordings from the soleus and tibialis anterior (TA) muscle were collected from the dominant leg. Measurements were obtained in three positions: while the participant was: 1) seated in a chair, hands resting in the lap, feet on the ground, and no volitional activation of target soleus muscle or TA muscle (sit rest); 2) seated in a chair, hands resting in the lap, feet on the ground, and asked to maintain a low-level volitional activation of target soleus muscle (sit active); and 3) in an upright stance with their feet shoulder-width apart, arms down at the side, and facing straight ahead (quiet stance) (Fig. 1).

Fig. 1.

Representative H-reflex traces from a young adult (YA) (left) and older adult (OA) (right) participant, respectively, during each task condition (sit rest, sit active, quiet stance) and at the three inter-stimulus intervals, or ISIs (−1.5 ms, +10 ms, +40 ms). Following the stimulus artifact, in each trace, the dashed black lines (YA) and dashed gray lines (OA) depict unconditioned (UC) H-reflex responses, while the solid, colored lines depict the conditioned H-reflex responses at each of the 3 ISIs (−1.5 ms = blue; +10 ms = red; +40 ms = purple). Note the slightly larger amplitude of the conditioned versus unconditioned H-reflex at the −1.5 ms ISI. At the + 10 ms ISI, the conditioned H-reflex is much larger than the unconditioned H-reflex, consistent with a large % modulation. At the + 40 ms ISI, in OA, compared to unconditioned H-reflexes, the conditioned H-reflex amplitude shows a decrease during task-related activation (sit active and quiet stance). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Electromyography (EMG) procedures

Surface EMG sensors (2-cm diameter, EL503 Biopac Systems Inc., Goleta, CA) were placed on the skin overlying the TA and soleus muscles, with a common ground electrode placed on the ipsilateral lateral malleolus. To verify proper EMG sensor placement, participants were asked to contract each muscle while an experimenter confirmed EMG activation (Biopac Systems AcqKnowledge Software Version 4.4). Next, EMG activity during a maximal voluntary contraction (MVC) of the TA and soleus was recorded. MVCs were completed through resisted dorsiflexion of the ankle in a seated position for the TA and standing single leg heel raises for the soleus. The participants were asked to maintain maximal contraction of the target muscle for 5 s for 2 trials, after which the MVC was determined using the largest rectified EMG amplitude produced.

Visual biofeedback regarding ongoing TA and soleus EMG (LabVIEW, National Instruments Corporation) was used to enable participants to maintain consistent background EMG activation during the experimental session, and for the experimenters to ensure target level of muscle activation remained consistent during testing. During the seated active task condition, participants were asked to maintain EMG activation equivalent to 10 % of their maximal EMG amplitude obtained during the soleus MVC, which is a low-level activation similar in magnitude to activation during upright quiet stance. EMG activity during quiet stance was verified by having the participant stand quietly for 30 s, during which the TA exhibited little to no muscle activity and the soleus slightly activated, with visual feedback display of ongoing soleus and TA EMG available to both participant and experimenter, similar to the 10 % soleus EMG activation during the seated active task condition. The experimenters monitored background EMG activity from both the targeted (Soleus, ankle plantarflexor) and the antagonist (tibialis anterior, ankle dorsiflexor) muscles throughout the experiment. By monitoring EMG activity in both the ankle muscles, we aimed to prevent variations in plantar- or dorsi-flexor muscle activation during quiet standing or other task conditions from affecting the H-reflex data. Notably, the experimenter monitored and provided feedback to the participants regarding their posture and limb positioning, especially during the quiet standing task, to maintain consistency during the experiment.

Peripheral nerve stimulation (PNS) procedures

To elicit soleus Hoffmann (H-) reflexes, peripheral nerve stimulation (PNS) was applied to the posterior tibial nerve within the popliteal fossa of the dominant lower limb using a monopolar cathode (round, 1-cm diameter), with the anode (square, 2 cm X 2 cm) placed at the midline proximal to the patella (Biopac Systems Inc, Goleta, CA). The stimulating electrodes were self-adhering carbon rubber electrodes. To find the optimal PNS site, the participant was asked to remain in a quiet standing position with equal weight bearing of bilateral lower extremities. The optimal location of the cathode was determined by sending PNS pulses and noting the location that elicited the largest H-reflex amplitude and evoked contraction of the soleus muscle. Once optimal cathode location was determined, inelastic overwrap was used to secure the cathode in place. EMG sensors and stimulating electrodes were also secured to the skin with tape to ensure maximal contact.

Next, the PNS intensities required for the H-reflex threshold and the maximal M−wave response (Mmax) were obtained during each task condition: sit rest (SR), sit active (SA), and quiet stance (QS). The H-reflex threshold was defined as the lowest PNS intensity where an H-reflex response greater than 0.1 mV was visible. The intensity for Mmax was determined after three increasing PNS intensities were sent, and the M−wave response plateaued. Soleus H-reflex and M−wave recruitment curves were collected during each task condition (SR, SA, QS). The delivery of increasing PNS intensities used for the 60 pulses was determined by calculating 20 % below the intensity needed to reach H-threshold and 20 % above the intensity needed to elicit Mmax; then, the difference between these calculated numbers was divided by 60, representative of the 60 pulses delivered during H-reflex recruitment curve collection For collection of the H-reflex recruitment curve, the PNS pulses were delivered incrementally in increasing order [23]. Following these 60 pulses, a H-reflex recruitment curve was plotted, and intensity required to elicit 50 % of the maximum H-reflex (H-max) amplitude was determined for use during TMS-conditioning for each task condition, respectively. An intensity that elicited an H-reflex amplitude equivalent to 50 % of H-max was used consistent with our previous study to allow sufficient range for upward and downward modulation of the conditioned H-reflex responses [26].

Transcranial magnetic stimulation (TMS) procedures

Stereotaxic neuronavigation (Brainsight v. 2.2.14, Rogue Research Inc., Canada) was used to register and scale the participant’s head to a standard brain template, and track coil location with respect to the TMS stimulation site. During TMS evaluation, the participant’s posture, position of lower and upper limb joints, and background activation state of soleus and TA muscles were monitored by the experimenters to maintain consistency throughout the experiment. Motor evoked potentials (MEPs) in the target soleus muscle were recorded in response to TMS, using surface EMG sensors. TMS pulses were delivered through a Magstim 200² stimulator using a custom 70 mm figure-of-eight batwing coil (Magstim Company Ltd., Dyfed, UK). Soleus hotspot was determined as the location on the primary motor cortex (M1) that elicited the largest MEPs from the soleus muscle. Stereotactic neuronavigation was used to ensure consistent TMS delivery over identified hotspot and accurate coil positioning during the experimental session. Resting motor threshold (rMT) and active motor threshold (aMT) were determined by stimulating over the M1 hotspot to elicit soleus MEPs. rMT was determined as the lowest TMS intensity that elicited ≥ 3/5 MEPs of an amplitude greater than 50 μV. The aMT was the lowest TMS intensity that elicited ≥ 3/5 MEPs of amplitude greater than 100 μV. After determining the TMS thresholds, to confirm the threshold and for ancillary analysis, we collected MEP data in response to trials at subthreshold (90 % MT) and suprathreshold intensities (120 % MT).

TMS-conditioning of soleus H-reflexes: methods and rationale

To investigate the influence of descending pathways on spinal reflex excitability, a PNS intensity that elicited an H-reflex amplitude of 50 % H-max was paired with a subthreshold (90 % rMT or aMT) TMS pulses at three different inter-stimulus intervals, or ISIs (−1.5 ms, +10 ms, +40 ms). The negative ISI (−1.5 ms) indicates that the PNS pulse was sent prior to the subthreshold TMS pulse, and the two positive ISIs (+10 ms, +40 ms) indicate the PNS pulse was sent after the subthreshold TMS pulse. TMS-conditioned soleus H-reflexes were collected during each of the 3 task conditions (SR, SA, and QS) (Fig. 1).

The methodological approach of pairing non-invasive stimulation of descending cortical motor (with TMS) and ascending spinal afferent pathways (using PNS) helps provide an indirect readout of the strength of descending influences on the H-reflex. By comparing the size of the unconditioned (PNS only) versus conditioned (i.e. PNS paired with TMS) H-reflexes, we can infer the magnitude and direction (i.e. facilitation versus inhibition) of descending modulation of spinal reflex circuits. Because we keep the stimulation parameters and background activity identical between the unconditioned versus conditioned H-reflex responses, when the conditioned H-reflex amplitude increases or decreases, this provides evidence of facilitatory or inhibitory, respectively, influences of the descending TMS-induced volleys on the local spinal reflex circuit. For the −1.5 ms ISI, when the PNS pulse is delivered 1.5 ms before the TMS pulse, the TMS-induced volleys can activate descending pathways that modify the excitability of the LMN pool, but with a very short time course, which is purported to be mediated by fast-conducting, largely direct or monosynaptic corticospinal tract neurons. In contrast, at the longer ISIs (+10 ms and + 40 ms), the TMS-induced conditioning volley can arrive at the spinal segmental level much more in advance of the PNS-induced test volley. Polysynaptic and/or slower-conducting corticofugal descending pathways can exert their influence and induce a change in the LMN pool excitability at these longer ISIs (+10 ms, +40 ms) [26,27,46,51]. In the current study, to minimize participant burden, fatigue, and data-collection duration, we utilized 3 standardized ISIs to measure TMS-conditioning of H-reflexes at one short (−1.5 ms) and two longer intervals (+10 ms, +40 ms), using parameters consistent with previous published methods [23,46].

Prior to each session, the order of each task condition and the order of the ISIs were randomized in Microsoft Excel using the RAND() function. Five blocks of stimuli sent approximately 5–6 s apart were delivered in the following order: PNS alone to elicit an unconditioned (UC) soleus H-reflex, followed by 3 sets of paired PNS + TMS pulses to condition soleus H-reflexes at each of the 3 ISIs (i.e., −1.5 ms, +10 ms, +40 ms). The main outcome to measure the magnitude of TMS conditioning of soleus H-reflexes was the Conditioned H-reflex %, calculated using the peak-to-peak amplitude of the conditioned H-reflex response compared to the peak-to-peak amplitude of the unconditioned H-reflex response x 100 % (Fig. 1).

Beam-walking task

The Narrowing Beam Walking Test (NBWT) is a performance-based clinical test of dynamic balance and gait function, developed to challenge balance control by constraining step width and/or reducing the support surface [48]. The NBWT is relatively safe and easy to administer, and has been shown to provide a sensitive measure of balance performance in neurologically unimpaired individuals across the lifespan [48]. We used the NBWT as a behavioral outcome in our study because it is one of the few clinical tests that do not show a ceiling effect in young or older people without neurological impairments, is relatively quick to administer, is clinically-relevant, and shows inter-individual variability in scores [48,52]. Consistent with methods used in previous studies, participants were asked to cross their arms over their trunk and walk along the narrowing beam while wearing standardized shoes [48,52]. Each trial concluded when the participant 1) stepped off the beam, 2) uncrossed their arms, or 3) walked the full length of the beam. The distance walked, divided by the total length of the beam (6.71 m), was recorded as the normalized distance. The average normalized distance over trials 3 to 5 was used as the NBWT score.

Statistical analyses

Normality was assessed using quantile–quantile (Q-Q) plots. Sphericity was assessed using Mauchly’s test, and if sphericity could not be assumed, the Greenhouse-Geisser correction was applied to the degrees of freedom. The study sample size was estimated based on the magnitude of effects observed in young adults in our previous study as well as our preliminary dataset on older adults [23]. Significant main effects were followed by post hoc contrasts and corrected for multiple comparisons using the Bonferroni method. Due to some participants not completing all task conditions (e.g., several missing samples in the datasets), restricted maximum likelihood (REML) calculations were performed for mixed-effects model analysis. First, to evaluate effects in the SR condition, a two-way repeated measures analysis of variance (ANOVA) was performed to evaluate the effects of timing, or ISIs (−1.5 ms, +10 ms, +40 ms), and aging (YA, OA) on TMS-conditioned soleus H-reflexes during SR. Next, to characterize the effects of task-related activation (SR, SA, QS) and aging (YA, OA) on TMS-conditioned soleus H-reflexes, mixed-effects analyses were performed at each ISI (−1.5 ms, +10 ms, +40 ms). Additionally, at the + 40 ms ISI during SA, a Pearson correlation coefficient was computed to assess the relationship between age and % Conditioned H-reflex. Mixed-effects analyses were completed to compare the change in magnitude of the Conditioned H-reflex % from SA to SR and QS to SR for both groups (YA, OA) and at each ISI (−1.5 ms, +10 ms, +40 ms). Finally, Welch’s t test was performed to compare NBWT scores between young and older adult groups, and Pearson correlation coefficients were calculated to assess the linear relationship between NBWT score and Conditioned H-reflex % at the + 40 ms ISI. All statistical tests were run in GraphPad Prism version 9.4 and the critical alpha level was set to p < 0.05.

Results

Complete datasets for all 3 task conditions were collected for 20 of the 21 young adult (YA) participants and 13 of the 26 older adult (OA) participants. The reasons for incomplete data included some participants completing only some experimental conditions during their session due to data-collection duration or scheduling (N = 10 OA), and data not being usable for analysis due to high variability or inconsistency in the H-reflex responses (N = 1 YA, N = 3 OA).

Effects of timing (ISI) and aging on TMS-conditioned soleus H-reflexes at rest

The two-way repeated measures ANOVA revealed a significant main effect of timing on the Conditioned H-reflex % at rest [F_{(1.594, 60.57)} = 37.57, p < 0.0001], but no significant main effect of aging [F_{(1, 38)} = 0.065, p = 0.7999). Post hoc pairwise comparisons pooled across groups revealed significantly greater Conditioned H-reflex % at the + 10 ms ISI compared to the −1.5 ms ISI (p < 0.0001) and the + 40 ms ISI (p < 0.0001), and significantly different Conditioned H-reflex % at the −1.5 ms ISI compared to the + 40 ms ISI (p = 0.03) (Fig. 2).

Fig. 2.

Conditioned H-reflex % at each ISI during the sit rest (SR) task condition. During SR task condition, both young adult (YA) and older adult (OA) groups showed significantly greater magnitude of Conditioned H-reflex % at the + 10 ms ISI compared to the −1.5 ms ISI (p < 0.0001) and the + 40 ms ISI (p < 0.0001). Error bars depict standard error of the mean. Additionally, the Conditioned H-reflex % at the 40 ms ISI was significantly greater than the −1.5 ms ISI (p = 0.027). *p < 0.05; ****p < 0.0001.

Effects of task-related activation and aging on TMS-conditioned soleus H-reflexes at each ISI

To investigate the effects of task-related activation at the −1.5 ms ISI, mixed-effects analysis revealed a significant main effect of task condition on the Conditioned H-reflex % [F_{(1.761, 59.86)} = 7.953, p = 0.001], but no significant main effect of aging [F_{(1, 41)} = 3.584, p = 0.065]. Post hoc pairwise comparisons pooled across groups revealed a significantly larger magnitude of Conditioned H-reflex % during sit active (p = 0.009) and quiet stance (p < 0.0001) compared to sit rest (Fig. 3A).

Fig. 3.

Conditioned H-reflex % at the −1.5 ms and + 10 ms ISI during three task conditions (SR, SA, QS). (A) At the −1.5 ms ISI (direct, faster), the Conditioned H-reflex % was significantly greater during SA (p = 0.009) and QS (p < 0.0001) compared to SR. (B) At the + 10 ms ISI (indirect, slower), the Conditioned H-reflex % was significantly less during SA (p < 0.001) and QS (p = 0.007) compared to SR. **p < 0.01; ***p < 0.001; ****p < 0.0001. Error bars depict standard error of the mean.

Additionally, to investigate the effects of task-related activation and aging on the + 10 ms ISI, mixed-effects analysis revealed a significant main effect of task condition on the Conditioned H-reflex % [F_{(1.758, 96.71)} = 11.19, p < 0.0001], but no significant main effect of aging [F_{(1, 110)} = 2.911, p = 0.091]. Post hoc pairwise comparisons pooled across groups revealed significantly reduced magnitude of Conditioned H-reflex % during sit active (p = 0.0009) and quiet stance (p = 0.0073) compared to sit rest (Fig. 3B).

To investigate the effects of task-related activation and aging at the + 40 ms ISI, mixed-effects analysis revealed a significant main effect for task condition [F_{(1.775, 97.65)} = 19.12, p < 0.0001], aging [F_{(1, 110)} = 31.54, p < 0.0001], and an interaction between task condition and aging [F_{(2, 110)} = 7.992, p = 0.0006] (Fig. 4A). Post hoc pairwise comparisons revealed the OA group had significantly different magnitude of Conditioned H-reflex % during sit active (p < 0.0001) and quiet stance (p < 0.0001) compared to the YA group, with the direction of modulation reversing toward inhibition in OA (Fig. 4B). Additionally, only within the OA group, the magnitude of Conditioned H-reflex % was significantly different (with a larger magnitude of inhibition) during sit active (p = 0.0002) and quiet stance (p = 0.0004) compared to sit rest (Fig. 4C), while such task-related differences were not observed in the YA group (sit active vs. sit rest = p > 0.999; quiet stance vs. sit rest = p > 0.999). Finally, a Pearson correlation analysis revealed a significant negative correlation between Conditioned H-reflex % at the + 40 ms ISI during sit active and age in years (r = -0.470, p = 0.037) in the OA group, such that older individuals showed a greater magnitude of inhibition at the + 40 ms ISI (conditioned H-reflex amplitudes less than 100 %) (Fig. 4D).

Fig. 4.

Conditioned H-reflex % at the + 40 ms ISI during all three task conditions for both young adult (YA) and older adult (OA) groups. (A) Individual data points at the + 40 ms ISI (late indirect) for both YA and OA groups with the Conditioned H-reflex % on the y-axis and the three task conditions on the x-axis. (B) When compared to YA, OA showed significantly lower Conditioned H-reflex % at the + 40 ms ISI during the sit active (SA) (p < 0.0001) and quiet stance (QS) (p < 0.0001) task conditions. No significant difference was observed during sit rest (SR). (C) Within the OA group, the magnitude of Conditioned H-reflex % was significantly reduced during the sit active (SA) (p = 0.0002) and quiet stance (QS) (p = 0.0004) compared to the sit rest (SR) task condition. No significant differences were observed within the YA group. (D) When plotting the Conditioned H-reflex % as a function of age in years, a significant negative correlation was observed for the OA group, where the greater the age of an individual, the more reduced the Conditioned H-reflex % becomes. *p < 0.05; ***p < 0.001; ****p < 0.0001. Error bars depict standard error of the mean.

Change in Conditioned H-reflex % with task-related activation

To explore the change (Δ) in Conditioned H-reflex % between seated voluntary activation (sit active) versus resting while seated (sit rest) and across ISIs in both groups, mixed effects analysis revealed a significant main effect for ISI [F_{(1.661, 62.28)} = 17.91, p < 0.0001], no significant main effect for aging [F_{(1, 38)} = 0.002, p = 0.964], and an interaction between ISI and aging [F_{(2, 75)} = 4.021, p = 0.0219]. Post hoc pairwise comparisons revealed that Δ Conditioned H-reflex % (sit active minus sit rest) at the −1.5 ms ISI was significantly greater (more positive values indicating a larger increase in Conditioned H-reflex % in the SA task condition) compared to the + 10 ms ISI (p < 0.0001) and the + 40 ms ISI (p < 0.0001), with both ISIs showing negative values indicating reduced Conditioned H-reflex % in the sit active task condition. Additionally, at the + 40 ms ISI, Δ Conditioned H-reflex % (sit active minus sit rest) was significantly lower (more negative values indicating a greater decrease in Conditioned H-reflex % in the SA task condition) during sit active for the OA group compared to the YA group (p < 0.01) (Fig. 5A).

Fig. 5.

The change (Δ) Conditioned H-reflex % between task-related activation conditions (sit active (SA), quiet stance (QS)) versus sit rest (SR)) at each ISI, where the line at 0 indicates no change in the Conditioned H-reflex % between the task conditions. (A) Data for change from sit active (SA) to sit rest (SR) indicate that voluntary activation significantly decreased the magnitude of Conditioned H-reflex % at the + 10 ms (p < 0.0001) and + 40 ms (p < 0.0001) ISIs compared to the −1.5 ms ISI. Additionally, at the + 40 ms ISI, OA showed a greater decrease in the magnitude of Conditioned H-reflex % during voluntary activation compared to YA (p < 0.01). (B) Data for the change from quiet stance (QS) to sit rest (SR) indicate that postural activation significantly decreased the magnitude of Conditioned H-reflex % at the + 10 ms (p < 0.001) and + 40 ms (p < 0.0001) ISIs compared to the −1.5 ms ISI. *p < 0.05; ***p < 0.001; ****p < 0.0001.

To explore the change (Δ) in Conditioned H-reflex % between postural activation while in quiet stance versus resting while seated (sit rest) and across ISIs in both groups, mixed effects analysis revealed a significant main effect for ISI [F_{(1.455, 45.10)} = 14.10, p < 0.0001], no significant main effect for aging [F_{(1, 31)} = 2.88, p = 0.099], and a trend towards statistically significant interaction between ISI and aging [F_{(2, 62)} = 2.754, p = 0.071]. Post hoc pairwise comparisons revealed that when collapsed across groups, Δ Conditioned H-reflex % (quiet stance minus sit rest) at the −1.5 ms ISI was significantly greater (more positive values indicating a larger increase in Conditioned H-reflex % in the quiet stance task condition) during postural activation while in quiet stance compared to the + 10 ms ISI (p = 0.0003) and the + 40 ms ISI (p < 0.0001), which showed negative values indicating a decrease in Conditioned H-reflex % in the quiet stance task condition (Fig. 5B).

Dynamic balance ability and the relation to descending modulation of spinal reflexes

First, to confirm the effects of aging on dynamic balance ability (i.e., NBWT), a Welch’s t-test revealed that the older adult (OA) group performed significantly worse on the NBWT compared to the young adult (YA) group (p < 0.0001) (Fig. 6A). To evaluate a potential relationship between NBWT scores and Conditioned H-reflex % at the + 40 ms ISI, Pearson correlation analyses revealed no significant relationship during either the sit active (YA: p = 0.57; OA: p = 0.93) or the quiet stance task condition (YA: p = 0.38; OA: p = 0.32) for both groups (Fig. 6B).

Fig. 6.

Dynamic balance ability assessed via the narrowing beam walking test (NBWT) and the relation to Conditioned H-reflex % during task-related activation at the + 40 ms ISI. (A) The YA group performed significantly better on the NBWT assessment compared to the OA group (p < 0.0001). (B) Pearson correlation analyses revealed that at the + 40 ms ISI, both YA and OA groups showed no relationship between Conditioned H-reflex (%) and NBWT Score during task-related activation (i.e., SA and QS). Error bars depict standard error of the mean.

Discussion

Our study results revealed interesting interactions between the effects of aging and task-related activation on descending modulation of spinal reflexes. Our most important findings were observed at the + 40 ms ISI, where we observed a significant effect of age, as well as a significant interaction between age and task condition. At the + 40 ms ISI, during tasks involving both voluntary (i.e., in the sit active condition) and postural soleus muscle activation (i.e., during quiet stance), conditioned soleus H-reflexes showed a significant effect of aging, such that older adults showed a reversal of direction of H-reflex modulation from positive (facilitation) to negative (inhibition) (Fig. 4, Fig. 4). When characterizing the task-induced change in the magnitude of descending modulation of H-reflexes (i.e., sit active or quiet stance versus sit rest), at the + 40 ms ISI, older adults showed a larger task-related H-reflex modulation (sit active versus sit rest) (Fig. 5). The effect of aging on H-reflex modulation at the + 40 ms ISI was further supported by a significant negative correlation between participant age and magnitude of Conditioned H-reflex % at the + 40 ms ISI, with older individuals showing greater inhibition of the soleus H-reflex (Fig. 6). Taken together, our results showed that only in older adults, only at the + 40 ms ISI, and only in the active conditions (SA, QS), we observed an inhibition of the conditioned H reflex.

Aging and task-related activation influence TMS-conditioned soleus H-reflexes

By combining PNS with a subthreshold TMS pulse using a very short timing interval (i.e., −1.5 ms ISI), previous studies have probed direct, faster-conducting TMS-induced descending volleys that arrive at the spinal cord LMN pool prior to the arrival of the PNS-induced afferent signal. Alternatively, when combining PNS with subthreshold TMS pulse using longer timing intervals (i.e., +10 ms, +40 ms ISIs), we additionally allow for slower-conducting, indirect TMS-induced descending volleys to arrive at the spinal segmental level prior to the PNS-induced afferent signal. Our findings revealed that during a seated rest (SR) task condition, the effect of ISI (i.e., relative timing of TMS with respect to PNS) was seen across both groups (YA and OA), with no significant effect of aging on the descending modulation of spinal reflexes. As expected, the ISI differentially impacted the magnitude of the TMS-conditioned soleus H-reflex response. Our results for the SR condition showed that the largest magnitude of TMS-conditioned soleus H-reflexes occurred at the + 10 ms ISI, which aligns with previous findings in younger adults [23,26], and suggests that at this ISI (i.e., +10 ms), a large proportion of excitatory descending inputs can summate, increasing the excitability of the spinal LMN pool, and contribute to a greater magnitude of facilitation of the soleus H-reflex responses [26]. TMS-conditioning of H-reflexes at the + 10 ms ISI is potentially mediated by combined contributions of descending volleys in several different corticofugal pathways (e.g. cortico-reticulo-spinal, cortico-vestibulo-spinal) to temporal summation of excitatory post-synaptic potentials at the LMN membrane [[21], [22], [23], [24], [25], [26]].

In contrast to sit rest, during task-related activation, whether voluntary or postural, we found differential effects on TMS-conditioned soleus H-reflexes at different ISIs. At the 1.5 ms ISI, both young and older adult groups demonstrated increased magnitude of conditioned H-reflex modulation during task-related activation compared to the SR task condition, which suggests that regardless of the motor task, voluntary (SA) and postural (QS) activation may preferentially recruit direct, faster descending pathways (mostly through the corticospinal tract) that induce a facilitation of the TMS-conditioned soleus H-reflex responses. In contrast, at the + 10 ms ISI, both groups demonstrated reduced magnitude of modulation of soleus H-reflexes during task-related activation (SA and QS) compared to the SR task condition, which suggests that voluntary and postural activation may be accompanied by inhibition or decreased engagement of indirect or slower descending pathways. Notably, the + 40 ms ISI was the only one to show a significant effect of age (H-reflex inhibition in OA in contrast to facilitation in YA), as well as an interaction between age and task condition. At the + 40 ms ISI, we observed a larger magnitude of inhibition during SA and QS, suggesting stronger descending modulation (albeit inhibitory) of spinal reflexes during the active task conditions. To our knowledge, one previous study similarly revealed that during upright quiet stance (i.e., postural activation), young adults showed an inhibition of the soleus H-reflex at similar longer timing intervals (i.e., +40 to + 60 ms ISIs) [44].

Potential physiological mechanisms of aging-related changes in TMS-conditioned soleus H-reflexes

During the sit rest task condition, no significant aging-related differences were observed in the TMS-conditioned soleus H-reflex responses at any timing interval. These findings suggest that while aging may adversely impact neural excitability or speed of transmission, in the resting state, both direct and indirect descending pathways maintain the ability to modulate spinal circuit activity regardless of age. Significant effects of aging were observed only at the + 40 ms ISI. At the + 40 ms ISI, the magnitude of conditioned H-reflexes differed between young versus older adults, but only during task-related activation. Older adults showed a larger magnitude of inhibition during the sit active and quiet stance task conditions versus the sit rest condition. Also, the magnitude of H-reflex modulation at the + 40 ms ISI was negatively correlated with age. Taken together, our findings of between-group differences and task-related differences observed only at the + 40 ms ISI, combined with the significant correlation between and H-reflex modulation, also only observed at the + 40 ms ISI, suggest that aging causes a change in the physiological pathways mediated by the + 40 ms ISI.

A few different physiological circuit mechanisms may underlie the TMS-conditioning of spinal reflexes at the + 40 ms ISI. Previous studies have suggested that TMS-conditioning of H-reflexes at longer-intervals may be mediated by slower, polysynaptic corticofugal pathways as well as propriospinal and spinal inter-segmental circuits [26,46]. One possible explanation is that aging-related changes at the + 40 ms ISI are representative of heightened descending inhibitory control of reflex gain during task-related activation, potentially mediated through Ia presynaptic inhibition [43,53,54] or propriospinal networks [55,56]. A previous study demonstrated the inhibition of knee extensor H-reflexes at long ISIs, and suggested that convergence of descending (i.e., TMS-induced) and peripheral (e.g. PNS-induced proprioceptive, cutaneous, and other afferent volleys) may activate non-reciprocal group I interneurons [13]. These interneurons may induce postsynaptic inhibition, decreasing the excitability of LMN pools, and resulting in H-reflex inhibition.

Descending corticospinal and other corticofugal and subcortical pathways (e.g. reticulospinal, vestibulospinal, propriospinal) as well as peripheral sensory afferent inputs all converge at the level of the spinal LMNs, allowing the LMN output to be inhibited or facilitated according to task requirements. Previous studies showed that subthreshold M1 electrical stimulation inhibits soleus H-reflexes at some ISIs, and subthreshold TMS suppresses ongoing EMG activity in both upper and lower limb muscles [[57], [58], [59]] suggesting that descending pathways have inhibitory influences on spinal reflexes. Potentially, an ‘unmasking’ of these inhibitory descending influences and/or local inhibitory spinal circuits occurs at the + 40 ms ISI, and underlies the inhibition of soleus H-reflexes observed in our study. Additionally, greater inhibition at + 40 ms in older adults may also be related to aging-related change in innervation ratios [60] or increase in heteronymous inhibition [61]. While these are hypothesized mechanisms, it is notable that TMS and PNS are artificial inputs, and a physiological interpretation of our results should be made with caution, and merit more in-depth future investigation. For instance, one promising future research direction to help parse out the candidate neural pathways may be to compare the effects of motor tasks or neuromodulation targeting corticospinal, brain stem, and spinal circuits on TMS-conditioning of H-reflexes.

Dynamic balance ability was not associated with descending modulation of spinal reflexes

Balance control involves static and dynamic conditions, which can be subdivided into four types: static steady-state balance (e.g., standing), dynamic steady-state balance (e.g., walking), proactive balance (e.g., anticipated slip on a wet floor), and reactive balance (e.g., tripping) [62]. While some previous studies have implied that these different types of balance were minimally associated with each other [63,64], others have suggested that the various types of balance were highly interlinked [65,66]. Here, in this study, the Narrowing Beam Walking Test (NBWT) [48] was used to assess dynamic balance ability by comparing the normalized distance walked in both young and older adult groups. Our results revealed that older adults traversed a shorter distance on the NBWT which is in line with several prior findings that have demonstrated aging-related decline in balance ability and postural control [[67], [68], [69]]. However, given the lack of associations between balance and neurophysiologic variables in our current study, the interactions between both cortical and spinal sensorimotor circuits that are responsible for aging-related declines in balance ability remain poorly understood.

Previous studies have shown that aging resulted in greater cortical engagement measured using electroencephalography (EEG) signals during performance of balance tasks, which appears to be a compensatory mechanism for older adults with lower balance ability [52,69,70]. Additionally, the soleus H-reflex response has been shown to be sensitive to postural changes in upright balance tasks [71,72], as well as during balance perturbations [[73], [74], [75]]. Here, we provided the first concurrent assessment of dynamic balance ability and descending cortical modulation of spinal reflexes in the same participant cohort. Our findings revealed no significant relationship between dynamic balance ability and descending modulation of spinal reflexes for both young and older adult groups. However, there was an association between age and descending modulation of spinal reflexes. Future work should explore other measurements of reactive balance that may exhibit a stronger relationship with descending modulation of spinal reflexes.

Implications for future research

Evaluating the effects of task-related activation and aging-related changes on descending cortical modulation of spinal reflexes can reveal insights into the underlying neural mechanisms of descending cortical and spinal circuit interactions implicated in lower limb sensorimotor control, which may inform future investigations in individuals with neurologic impairments, such as stroke, spinal cord injury (SCI), or Parkinson’s disease. Circuit interactions between descending cortical pathways and spinal sensorimotor reflex circuits play a very important role in the control of voluntary movement, yet are poorly understood [6,76,77]. Current non-invasive neurophysiologic measures, such as standalone single-pulse TMS or H-reflex amplitudes, are unable to specifically evaluate the connections or interactions between cortical and spinal circuits in humans, leading to a gap in our understanding of a salient mechanism underlying recovery of movement associated with rehabilitative interventions. Additionally, evaluating both TMS-conditioning of spinal reflexes at short- and longer-intervals, similar to the current study, could determine whether rehabilitation interventions differentially modulate the excitability of different descending neuromotor pathways, an aspect of training-induced neuroplasticity that is not well-understood. Future studies are needed to examine the effects of neurologic conditions, such as stroke or Parkinson’s disease on TMS-conditioning of H-reflexes, to parse out aging-related changes from those caused by neurologic impairment. Our work may also provide valuable approaches and neurophysiologic outcome variables for studying rehabilitation-induced neuroplasticity of direct and indirect descending cortical modulation of spinal sensorimotor circuitry.

Study limitations

In the current study, the same three task conditions (SR, SA, QS) and standardized ISIs (−1.5 ms, +10 ms, +40 ms) were used for measurement of TMS-conditioned soleus H-reflex responses for each participant. The stimulation subthreshold TMS conditioning intensity and the PNS intensity were also consistent, determined for each task condition and for each individual at the beginning of each experimental session. Furthermore, the EMG sensor placement location (e.g., over the muscle belly of the TA and Soleus), the PNS stimulation site (e.g., posterior tibial nerve within the popliteal fossa of the participant’s dominant leg), and the TMS coil location (e.g., placed over the Soleus M1 ‘hotspot’) were kept the same for each individual across all 3 task conditions. Thus, our experimental design was such that any differences in the amplitude of conditioned soleus H-reflexes observed in the −1.5 ms, +10 ms, and + 40 ms ISIs compared to unconditioned soleus H-reflexes could not be ascribed to differences in PNS intensity, TMS intensity, or other methodological concerns, which enhances rigor.

However, the current study has limitations. Throughout the experimental session, we monitored and displayed to the participant real-time visual feedback regarding the ongoing background EMG activation with respect to the target EMG. However, we did not record center of pressure and postural sway during standing, which would be an important addition to future studies. The experimenters did monitor and provide participants feedback (if needed) regarding ongoing posture and limb positioning to maintain consistency during the experiment. The technique of TMS-conditioning of H-reflexes is limited to those muscles from which H-reflexes can be consistently elicited. Soleus and other lower limb postural muscles may also receive stronger spinal network inputs, which could have influenced our findings. Previous findings have indicated that some participants may perceive pain in response to a standardized protocol to elicit soleus H-reflexes [78,79] but participants in the current study did not report discomfort with study procedures. Additionally, while we utilized a standardized set of ISIs for all participants in this study, future work would benefit from delivering customized ISIs based on each individual’s conduction velocities and physiological parameters, especially in older adults with neuro-pathologies. We acknowledge that while aging may be an important factor influencing the results observed at the + 40 ms ISI, other potential methodological (e.g. individual variability in central and peripheral conduction velocities, fluctuations of background muscle activity) or subject-specific factors (e.g. aging-related comorbidities or sub-clinical manifestations that affect neuromechanics) may have contributed to our observed results, and could benefit from being parsed out in future studies. Finally, for female subjects, it is possible that menstrual cycle effects could have influenced cortical [80] and spinal reflex excitability [81].

Conclusions

The current study findings provided the first characterization of aging-related effects on descending modulation of spinal sensorimotor reflex circuits under different postural and muscle activity states. Our findings of between-group differences in TMS-conditioning of H-reflexes at the + 40 ms ISI and correlations with age for TMS-conditioning of soleus H-reflexes at the + 40 ms ISI suggest that aging-related changes differentially affect how indirect slower-conducting pathways interact with spinal reflex circuits, which merits more investigation in future studies. Additionally, our findings further elucidate the need to study how aging affects the complex cortical, subcortical, and spinal circuit influences on the spinal segmental reflexes and LMN pool. By identifying aging-related changes in sensorimotor circuit function that were task-specific, these current findings may contribute to a better mechanistic understanding of aging-related deficits in sensorimotor circuit function, and inform future targeted intervention and neuromodulation strategies.

CRediT authorship contribution statement

Alejandro J. Lopez: Writing – review & editing, Writing – original draft, Project administration, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Terrence J. Glover: Data curation, Methodology, Software, Validation, Writing – review & editing. Taylor M. Leone: Writing – review & editing, Methodology, Formal analysis, Data curation. Catherine F. Mason: Writing – review & editing, Methodology, Data curation. Camille Guzman: Writing – review & editing, Writing – original draft, Methodology, Formal analysis. Lena H. Ting: Funding acquisition, Resources, Writing – review & editing. Michael R. Borich: Writing – review & editing, Writing – original draft, Visualization, Supervision, Resources, Project administration, Methodology, Investigation, Funding acquisition, Formal analysis, Data curation, Conceptualization. Trisha M. Kesar: Writing – review & editing, Writing – original draft, Visualization, Validation, Supervision, Resources, Project administration, Methodology, Investigation, Funding acquisition, Formal analysis, Data curation, Conceptualization.

Funding

NIH NICHD grants R01HD095975 (Kesar), K01HD079584 (Kesar), R01HD046922 (Ting).

Declaration of competing interest

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