Spinal Reflexes and Reflex Arcs

A reflex is an involuntary, stereotyped motor response to a specific sensory stimulus, mediated by a fixed anatomical circuit called a reflex arc. The arc always contains a receptor that transduces a stimulus into an afferent (sensory) signal, a sensory neuron whose axon enters the spinal cord, an integration center (the spinal cord gray matter itself, sometimes with interneurons), a motor neuron whose axon exits as the efferent limb, and an effector — a muscle or gland — that executes the response. Because the integration happens in the spinal cord (or brainstem, for cranial reflexes) rather than in cortex, reflexes can occur even when the spinal cord is severed from the brain, and they operate faster than any response requiring conscious perception. The single most important anatomical variable is synapse count. A monosynaptic reflex connects the afferent neuron directly onto the alpha motor neuron with one synapse in between — the stretch reflex (myotatic reflex) is the only true example in humans. A polysynaptic reflex interposes one or more spinal interneurons between afferent and efferent neurons, as in the withdrawal (flexor) reflex, the crossed-extensor reflex, and the Golgi tendon reflex (inverse myotatic reflex). Extra synapses cost time — roughly 0.5–1 ms of synaptic delay per synapse — but they buy computational power: interneurons can amplify, inhibit, coordinate multiple muscle groups across several joints, and route signals to the opposite side of the body. A second organizing principle is reciprocal innervation: whenever a reflex arc excites a muscle (the agonist), a branch of the same afferent fiber, via an inhibitory interneuron, simultaneously suppresses the antagonist motor neuron pool. This is not an accident of wiring — it is a necessity. Contracting the quadriceps while the hamstrings also fire would create a co-contraction that fights itself and wastes energy or even injures the joint. Reciprocal innervation guarantees smooth, unopposed movement at the reflex level, and the same circuit motif reappears, elaborated, in every voluntary movement you make.

Step 1 — Transduction at the receptor. Every reflex starts with a specialized sensory ending converting a physical or chemical stimulus into a graded receptor potential. In the stretch reflex, the receptor is the muscle spindle, a fusiform stretch-sensitive organ embedded in skeletal muscle, whose group Ia afferent fiber fires in proportion to the rate and magnitude of muscle stretch. In the withdrawal reflex, the receptors are free nerve endings — nociceptors — activated by noxious mechanical, thermal, or chemical stimuli, signaling via A-delta and C fibers.

Step 2 — Afferent conduction. The sensory neuron's axon (its cell body sits in the dorsal root ganglion, outside the cord) enters the spinal cord through the dorsal root. Conduction velocity varies enormously by fiber type: Ia afferents from muscle spindles are the largest, most heavily myelinated fibers in the body (12–20 μm diameter, up to 120 m/s), built for speed because stretch reflexes must react before a muscle overstretches. Nociceptive A-delta fibers conduct at 5–30 m/s; unmyelinated C fibers crawl at under 2 m/s — appropriate for a "second pain" signal rather than a split-second reflex trigger, since fast withdrawal is actually driven mainly by A-delta input.

Step 3 — Central integration. Inside the spinal gray matter, the arc branches. For a monosynaptic reflex, the Ia afferent synapses directly, via glutamate release onto AMPA/NMDA receptors, on the alpha motor neuron innervating the same muscle the spindle sits in (homonymous excitation) and on synergists. For a polysynaptic reflex, the afferent instead (or in addition) synapses onto one or more interneurons in the intermediate gray matter. These interneurons can be excitatory or inhibitory, can project up or down several spinal segments via propriospinal tracts, and can cross the midline. This is where divergence (one afferent influencing many motor pools) and convergence (many afferents influencing one motor neuron) occur, and where reciprocal inhibition and crossed circuits are built.

Step 4 — Efferent output. The alpha motor neuron, the final common path, fires and its axon exits via the ventral root, travels in a peripheral nerve, and terminates at the neuromuscular junction.

Step 5 — Effector response. Acetylcholine release triggers an end-plate potential, muscle fiber depolarization, and contraction — the visible reflex movement (the knee kick, the hand jerking off a hot pan).

Reciprocal innervation in detail. The Ia afferent from a stretched muscle (say, the quadriceps) does three things simultaneously: (1) it excites quadriceps alpha motor neurons monosynaptically; (2) it excites an inhibitory interneuron called the Ia inhibitory interneuron, which releases glycine onto hamstring alpha motor neurons, hyperpolarizing them; (3) it may also feed collaterals to ascending pathways (spinocerebellar, dorsal column) that inform the cerebellum and cortex of limb state without those pathways being part of the reflex itself. The net result is one afferent volley producing a coordinated, opposed pair of outputs — contraction of the agonist and relaxation of the antagonist — in well under 30 milliseconds.

The withdrawal reflex and crossed-extension. Step on a tack with your right foot, and the reflex must accomplish two things: pull the injured foot up (flexion) and keep you from falling by stiffening the opposite leg (extension). The nociceptive afferent enters the cord and, through interneurons, produces ipsilateral flexor excitation with reciprocal ipsilateral extensor inhibition (so the injured leg flexes cleanly), while a separate set of commissural interneurons crosses the midline to produce contralateral extensor excitation with reciprocal contralateral flexor inhibition (so the standing leg locks and bears weight). This is the crossed-extensor reflex, and it is inseparably polysynaptic — it requires at minimum an ipsilateral interneuron and a contralateral (commissural) interneuron, adding synaptic delay but making possible a whole-body postural adjustment from a single local stimulus.

The concept of the reflex arc was not obvious; it had to be extracted from careful comparative and lesion experiments across more than a century.

Charles Sherrington's decerebrate and spinal preparations (1890s–1900s). Sherrington, working mainly on dogs and cats, transected the spinal cord or brainstem at defined levels and studied the movements that survived. In a spinal animal (cord severed from the brain), stimulating the skin of a paw still produced a coordinated flexion-withdrawal of that limb plus extension of the opposite limb — proof that the circuitry for this coordinated response resides entirely within the spinal cord, not the brain. Sherrington used the term "final common path" for the motor neuron (building on which this article's prerequisite article is named) and demonstrated reciprocal innervation directly: electrically stimulating an afferent nerve that excited flexor motor neurons could be shown, with intracellular-style recordings of muscle tension and later with microelectrodes, to simultaneously reduce the tone or discharge of antagonist extensor motor neurons. He formalized this as "Sherrington's Law of Reciprocal Innervation." For this body of work on the integrative action of the nervous system he shared the Nobel Prize in Physiology or Medicine in 1932 with Edgar Adrian.

Demonstrating monosynapticity — Lloyd and the latency argument (1943). The claim that the stretch reflex involves exactly one central synapse required more than anatomy; it required a quantitative argument. D.P.C. Lloyd, working in Sherrington's intellectual lineage at the Rockefeller Institute, measured the central delay of the reflex — the time between the arrival of the afferent volley at the spinal cord and the appearance of the efferent discharge — by recording from dorsal and ventral roots. He found a delay of about 0.5–1.0 millisecond, matching the delay known for a single chemical synapse (from separate experiments on known monosynaptic-like preparations), and far shorter than the multiples of that delay seen in reflexes known to involve several interneurons (like the flexor withdrawal reflex, whose central delay was several milliseconds, consistent with 2–4 sequential synapses). This latency-counting method — central delay divided by known single-synapse delay — is still the standard way electrophysiologists infer synapse number in a circuit that cannot be directly visualized synapse-by-synapse.

The H-reflex as a human electrophysiological probe (Hoffmann, 1918; Magladery and colleagues, 1950s). Paul Hoffmann discovered that a low-intensity electrical shock to a mixed nerve (e.g., the tibial nerve) preferentially activates the large, low-threshold Ia sensory afferents before it activates motor axons, evoking a delayed muscle response called the H-reflex — the electrical analogue of the mechanically evoked tendon jerk. Because the H-reflex uses electrical stimulation instead of a tendon tap, it bypasses the muscle spindle's mechanical transduction step and allows extremely precise measurement of conduction velocities and central delay in intact humans. Comparing the H-reflex latency to the direct motor response (the "M-wave," from direct stimulation of motor axons) let researchers calculate conduction velocities of Ia afferents in living human subjects and confirmed, in humans, the same monosynaptic timing Lloyd had established in animals. This test remains a standard clinical and research tool for assessing spinal reflex excitability today.

Convergence and divergence — intracellular recording (Eccles, 1950s). John Eccles used intracellular microelectrodes in motor neurons to record excitatory postsynaptic potentials (EPSPs) evoked by stimulating single or grouped afferent fibers. He showed that a single Ia afferent produces a small, graded EPSP in its homonymous motor neuron (monosynaptic), while a compound afferent volley recruits inhibitory postsynaptic potentials (IPSPs) in antagonist motor neurons with a slightly longer latency — direct electrophysiological confirmation that reciprocal inhibition requires one additional synapse (the Ia inhibitory interneuron), exactly as Sherrington's behavioral experiments had implied decades earlier. Eccles shared the 1963 Nobel Prize in Physiology or Medicine (with Hodgkin and Huxley) largely for this synaptic-level dissection of excitation and inhibition in the spinal cord. Together these experiments moved the reflex arc from a descriptive anatomical sketch to a quantitatively verified circuit: a fixed number of synapses, a measurable central delay, and directly recorded excitatory and inhibitory postsynaptic potentials matching the predicted wiring diagram.

Example 1 — Timing a reflex and counting synapses. A clinical neurophysiology student stimulates the tibial nerve at the popliteal fossa and records from the soleus muscle. She measures:

• Distance from stimulation site to spinal cord entry (dorsal root): 0.50 m
• Distance from spinal cord (ventral root exit) to soleus muscle: 0.55 m
• Conduction velocity of the Ia afferent: 80 m/s
• Conduction velocity of the alpha motor axon: 70 m/s
• Total measured latency of the H-reflex (stimulus to first muscle response): 32 ms
• Neuromuscular junction transmission + muscle fiber conduction to the recording electrode: roughly 2 ms (given)

Step through the timing budget: Afferent conduction time = 0.50 m ÷ 80 m/s = 6.25 ms Efferent conduction time = 0.55 m ÷ 70 m/s = 7.86 ms NMJ + muscle delay = 2 ms (given) Sum of known conduction delays = 6.25 + 7.86 + 2 = 16.1 ms Remaining, unaccounted time = 32 − 16.1 = 15.9 ms This residual is far larger than the ~1 ms expected for a single monosynaptic delay — it is roughly 15–16 times that value. The student concludes this cannot be a pure monosynaptic H-reflex latency; either her distance or velocity estimates are off, or (more likely, given the numbers she was handed) she is actually observing a polysynaptic component layered onto the reflex, or her stimulation intensity recruited slower afferents and additional interneuronal relays (as can happen with the flexor reflex afferents that share the same nerve trunk at higher stimulus intensities). The lesson: quantitative latency accounting is exactly how physiologists distinguish a clean monosynaptic response from a contaminated or genuinely polysynaptic one — precisely Lloyd's method, applied here as a diagnostic check rather than taken on faith.

Example 2 — A clinical reasoning case: absent knee jerk with intact withdrawal. A 58-year-old patient presents with an absent patellar (knee-jerk) reflex on the left side, but a normal, brisk withdrawal reflex when the left sole is stroked (plantar/nociceptive withdrawal is intact), and normal strength on manual muscle testing of the quadriceps. Reasoning: the patellar reflex is monosynaptic, requiring an intact chain of (1) muscle spindle, (2) Ia afferent fiber, (3) dorsal root entry, (4) a single synapse onto the quadriceps alpha motor neuron, (5) ventral root, (6) femoral nerve, (7) neuromuscular junction, (8) muscle. The withdrawal reflex uses a different, polysynaptic circuit relying on nociceptive A-delta afferents entering via different dorsal root levels and different interneuron pools, with efferents traveling in different nerves (e.g., peroneal/tibial rather than femoral) to different muscles (flexors, not quadriceps). Since withdrawal is intact, the ventral horn motor neurons and the general spinal segment are functioning, and gross muscle strength is preserved, ruling out a diffuse lower motor neuron lesion or primary myopathy at that level. The selective loss of only the monosynaptic component points toward a lesion specific to the large-diameter, heavily myelinated Ia afferent pathway — for example, a sensory (dorsal root) ganglionopathy, a proximal femoral neuropathy, or, classically, tabes dorsalis (posterior column and dorsal root degeneration from tertiary neurosyphilis), all of which selectively damage large myelinated proprioceptive fibers while sparing thinner nociceptive fibers and general motor axons. This dissociation — reflex loss with preserved strength and preserved pain-withdrawal — is a textbook localizing sign precisely because monosynaptic and polysynaptic reflexes use anatomically and physiologically distinct fiber populations, exactly the distinction this article has been building.

Sherrington trained in medicine at Cambridge and spent his early career doing painstaking, patient experimental physiology on the spinal cords of anesthetized and decerebrate animals at a time when neurons themselves were still debated entities (the neuron doctrine versus reticular theory controversy was ongoing). He coined the word "synapse" (at the suggestion of the classicist Arthur Verrall, working from the Greek for "clasp") to name the junction he inferred must exist between neurons based purely on the physiological behavior of reflexes — delay, one-way conduction, fatigue, and facilitation — decades before anyone could see a synapse under a microscope. His 1906 book, The Integrative Action of the Nervous System, synthesized a career's worth of reflex experiments into a coherent theory: the nervous system as a hierarchy of integrative levels, with the spinal reflex arc as its most basic, most thoroughly characterizable unit, and the motor neuron as the "final common path" onto which countless reflex and voluntary signals converge and compete.

Clinical neurology. The deep tendon reflex examination (biceps, triceps, brachioradialis, patellar, Achilles) remains one of the fastest, cheapest, most information-dense parts of the neurological exam. Hyporeflexia or areflexia suggests a lower motor neuron, peripheral nerve, or dorsal root problem (e.g., Guillain-Barré syndrome, radiculopathy, tabes dorsalis); hyperreflexia with clonus suggests an upper motor neuron lesion that has removed descending inhibitory control over the spinal reflex arc (e.g., stroke, spinal cord injury, multiple sclerosis). The Babinski sign — extension rather than flexion of the great toe on plantar stimulation — is a polysynaptic reflex whose pattern change (not presence or absence) localizes corticospinal tract damage.

Spinal cord injury and rehabilitation. After a complete spinal cord injury, reflex arcs caudal to the lesion survive (the circuitry is local) but lose descending modulation, producing spasticity, exaggerated withdrawal reflexes, and autonomic dysreflexia. Rehabilitation strategies, including partial body-weight-supported treadmill training, exploit surviving polysynaptic spinal circuitry — closely related to the central pattern generators previewed in Central Pattern Generators and Locomotion — to retrain stepping-like patterns even without full brain control.

Anesthesiology and intraoperative monitoring. The H-reflex and F-wave are used to monitor spinal cord and nerve root function during spine surgery, providing a real-time electrophysiological read on reflex arc integrity while the patient is unconscious.

Pharmacology. Drugs acting on spinal inhibitory interneurons (e.g., baclofen, a GABA-B agonist, and glycine-pathway–related agents) are first-line treatments for spasticity precisely because they restore or mimic the inhibitory half of reciprocal innervation that descending damage has disrupted.

Toxicology and pesticide/nerve-agent research. Reflex circuits, being simple, quantifiable, and heavily reliant on cholinergic neuromuscular transmission, are classic assay systems for testing acetylcholinesterase inhibitors (organophosphates, nerve agents) and their antidotes, since disrupted reflex latency and tetanic response are sensitive, measurable biomarkers of neuromuscular poisoning.

Comparative and evolutionary biology. Reflex arc organization is deeply conserved from lamprey to human, making it a favored model for studying how simple, evolutionarily ancient circuits (crossed-extension in swimming vertebrates, for instance) were elaborated over evolution into the more complex locomotor pattern generators and voluntary motor hierarchies of mammals — a throughline this article sets up for the article on central pattern generators.

Ask the tutor: How does reciprocal innervation in a simple spinal reflex differ from the more elaborate reciprocal circuits used during rhythmic locomotion?