Reflexes are functionally useful and meaningful when the excitability of their pathways and behaviors can be modulated ( Fig. 1 B). Normally, the excitability of spinal reflex pathways is modulated according to the task being performed. For example, the excitability of the soleus H-reflex (a partial electrical analog of the spinal short-latency stretch reflex, Fig. 1 A ) pathway decreases from standing to walking to running, ^,, which helps to prevent saturation of this pathway, and further modulated to more specialized skills such as ballet dancing, jumping, and kicking a ball (see Fig. 1 B) such that this pathway can function appropriately in each of these motor skills and behaviors. In people with chronic SCI, regardless of the activity performed, the excitability of this reflex pathway can be stuck high and not modulated ( Fig. 1 C), and therefore, the reflex could be detrimental and a part of problematic spasticity.

Spinal stretch reflex and H-reflex, and their modulation in people without injuries and those with spastic hyperreflexia due to chronic spinal cord injury (SCI). ( A ) Main pathway of the spinal short-latency stretch reflex and the H-reflex. Excitation of Ia (and large II) spindle afferents activates α-motoneurons innervating the same muscle and its synergists largely monosynaptically. If the afferents are excited by muscle stretch, the response is the stretch reflex. If they are stimulated electrically, the response is the H-reflex. The activity of this spinal pathway is strongly affected by excitatory ( red ) and inhibitory ( navy ) input of spinal and supraspinal origins. ( B ) In an individual with no injuries, the spinal reflex pathway ( center ) responsible for the soleus H-reflex participates in many motor behaviors, ranging from standing to walking to running to athletic skills such as ballet dancing, jumping, and kicking a ball. The gain of the reflex pathway (+ and − in the gray circle ) is modulated between the tasks such that the input from muscle spindle afferents contributes appropriately to soleus activity in each of these movement tasks. ( C ) In a person with spasticity due to SCI, task-dependent reflex gain modulation is often impaired, and the pathway could remain hyperactive regardless of the movement tasks, and therefore nonfunctional.

( Modified from Ref. )

During dynamic motion such as locomotion, these reflexes are also phase-dependently modulated. For instance, the soleus H-reflex and stretch reflex are phase-dependently modulated during walking, in accord with soleus motoneuron excitability level (expressed indirectly through the electromyography [EMG] activity), which is high in the midstance phase and little-to-none in the swing phase. ^, In individuals with SCI, soleus H-reflex and stretch reflex are unsuppressed or often large in the midlate swing phase ( Fig. 2 ). This supports the possibility that these reflexes, if not suppressed adequately, would contribute to inappropriate soleus activity in the swing phase and exacerbate foot drop.

Phase-dependent modulation of the soleus H-reflex and stretch reflex during walking in people without neurologic injuries and people with spasticity due to chronic incomplete SCI. ( A ) Examples of the soleus H-reflex during the midstance and late-swing phase of and walking in a participant without known neurologic injuries ( top, black ) and in a participant with chronic incomplete SCI ( bottom, blue ). Height, weight, and age are similar in these 2 individuals. Each trace is the average of 10 or greater EMG sweeps and the time of stimulation is indicated with an arrowhead in each panel. In the participant without injuries, the H-reflex is present in the midstance when the soleus is active but completely suppressed in the late-swing phase when the soleus is not active. In the participant with SCI, the H-reflex is not suppressed in the late-swing. ( B ) Amplitudes (mean ± SE) of the soleus EMG, H-reflex, and M1 (soleus short-latency) and M2 (soleus medium-latency) stretch reflexes during walking in participants with chronic incomplete SCI and participants without. For calculating the soleus EMG, greater than 100 unperturbed steps were averaged for each participant. For calculating reflex sizes, for each participant in each of 8 equal bins of the step cycle, 10 or greater responses were averaged together and normalized to the maximum M-wave (M _max ) amplitude measured during standing. Bins 1 to 4 correspond to the stance phase, bin 5 to the stance-swing transition, and bins 6 to 8 to the swing phase. Horizontal lines labeled with “S” in indicate the average maximum H-reflex (H _max ) amplitudes during standing. For all panels, statistically significant differences between the groups ( red ) and between the bins ( blue , for the SCI group only) by Student t-test with Bonferroni correction are indicated with asterisks .

( Modified from Ref. )

Clonus, which is often associated with spastic hyperreflexia ^,, (however, see also ref ), also impairs walking. As elicitation of stretch reflexes requires only a degree or 2 of joint rotation, when those reflexes are not suppressed adequately, ground contact that produces a small joint motion perturbation can elicit stretch reflexes, which in turn trigger clonus in individuals with spastic hyperreflexia. Clonus triggered in the late swing to swing-stance transition phase often continues into the stance phase, exaggerates breaking force, reduces forward propulsion, continues into smaller fused push-off bursts and propulsive force generation, ^, thereby contributing to slow gait. Notably, muscle stretch and joint rotation are not the only input sources that trigger clonus; during walking, nonnoxious cutaneous nerve stimulation that produces cutaneous sensation around the foot often triggers clonus in the plantarflexors of spastic individuals with chronic SCI (nonpublished observations by the authors), similarly to dorsiflexion perturbation or H-reflex eliciting posterior tibial nerve stimulation reported previously. The fact that different kinds of afferent input can trigger clonus may further suggest that to some extent, the problems are in the motoneuron (ie, the final common path for generating motor function and behaviors), multiple spinal interneurons that mediate between the afferents and motoneurons, and/or deprivation of appropriate descending input to the spinal neural circuitry in general.

Abnormal excitability modulation of spinal reflexes arising from muscle spindle afferent ^,,,,,, is likely be further exacerbated by abnormal reciprocal inhibition between the antagonists. ^,,,,, Reflex abnormalities are also present in other spinal pathways, such as those of Ib inhibition, recurrent inhibition, and cutaneous reflexes, and are often accompanied by changes in spinal motoneurons and interneurons. ^,,, These pathways may also interact with each other. ^,,,, Altogether, in the spinal cord of individuals with chronic SCI, the excitability modulation and function of multiple pathways become impaired, contributing to abnormal motoneuron firing ^,,,,, and motor control. ^,, Based on these facts, we may view spasticity in chronic SCI as impaired regulation of reflexive excitation of motoneurons and inadequate suppression of motoneuron firing, at least partly.

The presence or absence of a reflex indicates whether that specific reflex pathway is available to excite motoneurons in a specific context in which the reflex is measured. ^,, For instance, in individuals without injuries, the soleus H-reflex and stretch reflex are present during the midstance of walking, indicating that these reflex pathways are available to excite soleus motoneurons; whereas during the mid-to-late swing phase, those reflexes are almost completely absent, indicating that their pathways are not available as the means to excite the same motoneurons (see Fig. 2 ). ^, In spastic individuals with SCI, the H-reflex and stretch reflex are present in the late-swing phase and can be robust and trigger clonus that continues into the following stance and leads to a chain of events that impair gait. ^,, A reflex that cannot be turned off when it should be or a reflex whose excitatory/inhibitory action cannot be asserted when it is critical ^, would not be useful but detrimental in motor control. Reflexes measured at a specific phase of movement or during a specific task may therefore give insight into normal or abnormal neural control of movements in people with SCI and others.

In the traditional clinical setting, reflex assessments are often made at rest or made without monitoring prestimulus ongoing EMG activity or posture. During such “resting” reflex measurements, the subthreshold level of motoneuron pool’s excitability is unknown and uncontrolled. That is, from the surface EMG signal, we cannot know how close or far the motoneuron pool’s excitability is from the firing threshold, and therefore, the variability in subthreshold level of motoneuron pool’s excitability from one reflex measurement to another would be unknown. Thus, when a reflex is measured at rest, a large sample size or a large effect size is likely necessary to overcome such measurement variability. When measuring a reflex at rest, the stimulation rate (for postactivation depression ^,,, ), the agonist and antagonist muscles’ ongoing EMG activity (for motoneuron pool’s excitability and postsynaptic and presynaptic inhibition ^,,,,, ), and joint posture (for posture-dependent reflex gain modulation ^,,, ) should be carefully considered in the protocol development. Reflexes are influenced by many factors. At rest, the excitability variability in many spinal neurons including motoneurons is unknown/uncontrolled. Therefore, resting reflexes should be interpreted carefully in association with spasticity and neuromuscular impairments.

Abnormal reflex behaviors observed at rest, during static motor tasks, or during isolated joint motion in individuals with SCI ^,,,,,,,, have encouraged us to assume that spinal reflex abnormalities contribute to spastic movement disorders. ^, However, in people with spastic hemiparesis due to stroke, stretch reflexes that are hyperexcitable at rest or in passive conditions are near normal or even suppressed during active muscle contractions, ^, and antispastic medication does not necessarily improve motor function. ^,,, In spastic individuals with chronic SCI, the soleus stretch reflex excitability is elevated during sitting or lying in supine ^, ; the reflex gain and reflex stiffness do not decrease with muscle activation ; and those reflexes are abnormally large in the mid-to-late swing phase of walking (see Fig. 2 ), likely contributing to spastic gait disorders. These observations are reminder that how we measure the reflex matters. That is, we can determine whether a reflex functions abnormally and contributes to impaired motion, through its examination in the actual dynamic motion; a reflex characterization measure made at rest would not immediately explain how it functions. In sum, reflexes are useful neural probes of the CNS, as long as we measure them correctly for the intended measurement purposes and apply appropriate interpretation on the gathered observations.

Operant conditioning is a powerful method for modifying a behavior based on its consequences. ^,,,, Through trial and error, a subject learns to produce a behavior that is rewarded. Through repetition, a trained neural behavior becomes a habitual behavior. With operant conditioning of a spinal reflex, a subject is rewarded for producing a larger (with up-conditioning) or a smaller (with down-conditioning) reflex. Since the amplitude of a reflex reflects the excitability of that reflex pathway at the time of reflex elicitation, ^,,, with reflex operant conditioning, we can train the CNS to produce the neural activity that increases (up-conditioning) or decreases (down-conditioning) the excitability of the targeted reflex pathway. ^,,,, As this is done repeatedly and persistently, over time, the spinal reflex pathway (including spinal interneurons and motoneurons) goes through anatomic and physiologic changes ^{,,,,,,,,,,,,,,,,,} ; and the changes persist after conditioning ends. ^{,,,,,,,,,,,,,} Markedly different from conventional pharmacologic therapies, reflex operant conditioning does not disable certain connections (eg, botulinum toxins) or enhance general inhibition (eg, baclofen), either of which would limit muscle activation and contraction regardless of the motor functions of the person’s intention. When a reflex that underlies spasticity is targeted by down-conditioning, ^, that reflex’s excitability decreases but preserves motoneuron’s ability to fire in response to voluntary effort or when it is functionally appropriate (eg, midlate stance phase of walking ; Fig. 3 ), ^, and thereby helps to improve locomotion in SCI. ^{,,,,,,,,,,,,,}

Soleus H-reflexes in individuals with spastic hyperreflexia due to chronic incomplete SCI, in whom H-reflex size decreased significantly through down-conditioning and effects of H-reflex down-conditioning on locomotor EMG activity. ( A ) Average H-reflexes for a baseline session ( black dotted line ) and the last conditioning session ( blue solid line ) from a participant with SCI. ( B ) Average (±SE) H-reflex size for baseline, conditioning, and 1 and 3 month follow-up sessions. ( C ) Locomotor EMG in the soleus and tibialis anterior (TA) in absolute value before ( dashed ) and after ( solid, green ) down-conditioning in an individual whose H-reflex size decreased through down-conditioning. Soleus push-off burst was increased and TA’s cocontraction with the plantarflexors in the stance phase was decreased after conditioning (modified from Ref. ). Push-off burst is highlighted by a red oval.