Article Text

Download PDFPDF

Why are upper motor neuron signs difficult to elicit in amyotrophic lateral sclerosis?
  1. Michael Swash1,2
  1. 1Royal London Hospital, Barts and the London School of Medicine, Queen Mary University of London, UK
  2. 2Institute of Neuroscience, University of Lisbon, Portugal
  1. Correspondence to Professor Michael Swash, Royal London Hospital, Barts and the London School of Medicine, Queen Mary University of London, London EC2Y 8BL, UK; mswash{at}


It is often difficult to identify signs of upper motor neuron lesion in the limbs of patients with amyotrophic lateral sclerosis, in whom there is neurogenic muscle wasting of varying severity. The reasons for this are complex and not related simply to the degree of lower motor neuron muscle wasting but, rather, depend on the pathophysiological abnormalities that develop in response to damage to descending motor pathways and to motor neurons and interneurons in the ventral horns of the spinal cord. The different mechanisms underlying the clinical phenomenology of the functional motor defect in amyotrophic lateral sclerosis, that lead to difficulty in detecting classical upper motor neuron signs, are discussed.

  • Motor neuron disease
  • anterior horn cell disorder
  • amyotrophic lateral sclerosis
  • spasticity
  • spinal cord
  • neurophysiology
  • neurourology
  • muscle disease, neuropathy

Statistics from


The diagnosis of amyotrophic lateral sclerosis (ALS) depends on recognition of a combination of upper motor neuron (UMN) and lower motor neuron (LMN) signs in various combinations in several bodily regions, inexplicable by any other diagnosis. Formal criteria for the clinical diagnosis have been agreed by consensus,1 2 and have become widely accepted.3 4 However, there is a generally recognised need for earlier clinical diagnosis than these somewhat rigid criteria allow. Part of the difficulty in achieving diagnostic certainty early in the course of the disease arises from difficulty recognising UMN signs when there is partial denervation in the limb.5 6 For example, Ince et al7 found corticospinal tract (CST) degeneration at autopsy in 50% of patients clinically diagnosed as progressive muscular atrophy in whom corticospinal signs had not been recognised in life.

The classical signs of UMN lesions are usually stated as: weakness, increased tendon reflexes, spasticity and an extensor plantar response.8 However, when there is additional neurogenic weakness due to LMN lesion in the same limb the tendon reflexes are often not clearly increased, spasticity cannot be demonstrated, supposedly because of the LMN weakness and the plantar response may be difficult to elicit, a difficulty often ascribed to differential weakness in toe and ankle flexor and extensor muscles. The plantar response is extensor in only some 50% of patients with ALS.9 The typical signs of LMN lesions are: weakness, muscle wasting, fasciculation, reduced tone and absent or reduced tendon reflexes.8 It is relevant to note that these classical features of UMN and LMN lesions were defined for recognition of LMN or UMN lesions as separate syndromes, not for differentiation when there is a combination of these features in the same limb. However, why is it often so difficult to evaluate possible UMN signs in ALS?

The UMN syndrome

The different components of the UMN syndrome reflect different physiological abnormalities in the descending motor system,10 expressed by the intact LMN system, including the segmental cord motor system. In the lower limb, weakness due to corticospinal dysfunction is most marked in foot dorsiflexion and hip flexion and, in the upper limb, there is a characteristic disturbance of fine co-ordinated finger movement. Indeed, impaired fine co-ordinated distal movement and slowness of complex movement patterns are often self-reported by patients with UMN lesions. However, in clinical practice, the UMN lesion is almost never confined to the rather small number of corticospinal fibres, arising from cortical Betz cells11 but involves other descending pathways in the internal capsule and its caudal projections, including prefrontal and extrapyramidal projections and rubrospinal, reticulospinal and vestbulospinal pathways.11 The clinical features of pure corticospinal lesion have been studied experimentally in the macaque,12 and have also been studied in humans with surgical or disease-related lesions.11 13 When there is damage to other descending motor fibres from extra-Rolandic motor cortical areas, additional features become evident as, for example, in patients with stroke or other large cerebral lesions. These more extensive lesions cause more diffuse weakness in affected limbs, with a rigid spastic syndrome. In this familiar syndrome, spinal segments are released from descending inhibition of the CST, and from the intrinsic propriospinal oligosegmental pathways that transmit the signal for movement from higher centres to motor neurons.14 Propriospinal neurons in the cervical and lumbar enlargements integrate excitatory and inhibitory input from ascending and descending pathways, including group Ia and Ib spindle afferents, and tendon organ afferents, inhibitory cutaneous afferents, corticospinal fibres, and rubrospinal, reticulospinal, vestibulospinal and other descending inputs (figure 1). In the classical UMN syndrome there is hyperexcitability at spinal segmental level,10 causing increased resting tone, and a characteristic velocity-dependent, stretch-sensitive increase in tone in response to passive limb movement across a joint, with the clasp-knife phenomenon representing developing inhibition during the movement, that overcomes the stretch-induced excitation and causes sudden relaxation of the tested agonist muscle.

Figure 1

The descending corticospinal tract (CST)input to propriospinal neurons (PNs) in the cervical and lumbar enlargements, directing limb movements. Propriospinal neurons are also found in thoracic segments. Note that the direct corticospinal projection from cortical Betz cells represents <5% of fibres in the corticospinal tract.11 Other descending non-corticospinal projections (vestibulospinal, rubrospinal and tectospinal pathways — ‘others’ in the diagram) are important in modulating vestibular, cerebellar and sensory input to the propriospinal command motor systems for limb and truncal movement in the spinal cord. At the spinal segmental (not shown) level, sensory input from muscle spindle and Golgi tendon organ afferents, and from cutaneous afferents, modulate output from anterior horn cells (AHCs) through direct or interneuron connexions. The γ motor system functions in parallel with the γ motor system. Afferent axons from muscle receptors, and from cutaneous and joint receptors, project rostrally to high cervical, thalamic and cortical levels through posterior columns, spinothalamic tracts and spinocerebellar afferents.

Tendon reflexes in the classical UMN syndrome are typically increased (brisk), often with repetitive clonus, representing unconstrained activity in the feedback loop to hyperexcitable anterior horn cells from Ia afferents during sudden muscle stretch. Increased tendon reflexes and clonus are not necessarily associated with spasticity, but they are important signs indicating that the monosynaptic stretch reflex loop is intact and disinhibited.

The plantar response is a polysynaptic cutaneomuscular response of particular interest since the demonstration of contraction of ankle muscles in the ‘extensor plantar response’, representing an enhanced physiological flexor withdrawal response in UMN syndrome involves excitation of muscles from cutaneous stimulation that can no longer be excited by voluntary command from the cortex. All these features of UMN syndrome depend on the integrity of segmental propriospinal, anterior horn and interneuronal anatomy in the cord, in the context of release from descending input from rostral motor control systems. In long-standing spasticity, affected muscles develop increased stiffness due to disuse, fibrosis, and tendon and joint contractures that further modify the classical findings.15

The UMN syndrome in ALS

The clinical difficulty recognising UMN features in ALS is often ascribed to LMN weakness which, it is said, may obscure the typical pattern of corticospinal weakness. There is always an UMN lesion in ALS, and LMN dysfunction, required for a definite diagnosis and confirmed at autopsy.16 The UMN lesion includes degeneration of the CST and related descending pathways in the internal capsule, brainstem and spinal cord. There are pathological changes in layer V of the motor cortex, with loss of Betz cells, but also more widespread pathology in the anterior brain, involving deep frontal and temporal white matter and the corpus callosum,17 detectable in life by diffusion tensor MRI,18 and degeneration in motor nuclei of the basal ganglia.16 Such extensive pathological changes would be expected to cause the full UMN syndrome. Does the concomitant occurrence of LMN neurogenic wasting in limb muscles by itself account for the difficulty in recognising features of UMN lesions in ALS so often noted on clinical examination, or are there additional factors?

Loss of spinal interneurons and gamma motor neurons in ALS

In ALS, in addition to loss of functional α motor neurons that innervate striated muscle fibres, there are other abnormalities in the cord. In the anterior horns of the spinal segments, motor neurons are arranged such that neurons innervating axial muscles are located in medially placed longitudinal nuclei, and neurons innervating distal muscles, such as small hand muscles, are situated laterally.19 Each spinal nucleus innervating a muscle contains interneuronal connexions from related flexor and extensor muscles, including internuncial connexions from the opposite side of the cord, presynaptic Renshaw inhibition, descending propriospinal interneuronal connexions, and input from muscle spindle and tendon afferents (figure 1). In humans, the interneurons associated with the segmental α motor neurons are located in close proximity. Small γ motor neurons innervating muscle spindles are located in relation to their functional associations with the spinal motor nuclei representing the innervation of specific muscles. Histological studies of the cord in ALS have revealed that interneurons degenerate alongside loss of α motor neurons20 and that the γ and β innervation to muscle spindles is also lost in the disease, although the primary and secondary sensory innervation of muscle spindles is intact.21 The α–γ coactivation during voluntary movement22 therefore is no longer possible. Exactly what effect this might have on dynamic stretch reflexes during voluntary movement, and the generation of spasticity, is uncertain, but de-efferented spindles have been shown to continue to function normally to muscle stretch, tendon vibration and tendon percussion.10 The response to a dynamic tendon stretch is principally dependent on the level of segmental excitation of motoneurons innervating the tested muscle.10

Tendon reflexes in ALS

It is generally accepted that assessment of tendon reflexes is difficult when there is marked LMN weakness involving the muscles activated by the usually tested tendon reflexes. For example, when the quadriceps or soleus/gastrocnemius complex is weak, wasted and fasciculating, it is very difficult to assess a tendon reflex as more active than ‘present’, rather than increased, and much bedside discussion may ensue as to whether the reflex is considered as ‘increased’ relative to the degree of wasting and weakness in the muscle, as a consequence of the coincident LMN abnormality.

In ALS, destruction of the intrinsic functional motor neuron and interneuron connexions within the anterior horn, including descending excitatory motor signals from propriospinal projections onto the remnants of the motoneuronal and interneuronal machinery, a reduced corticospinal inhibitory projection, and reduced interneuronal and presynaptic inhibition, will result in rather variable segmental excitation of remaining functional motor neurons. This will cause tendon reflexes to be less hyperactive than might be expected than in patients with isolated cerebral lesions, in whom the segmental motor neuronal structure is intact. This reduced responsiveness also accounts for the frequent difficulty in eliciting repetitive clonus at the ankle in ALS.

The H reflex, a response elicited in a muscle after submaximal stimulation of its mixed nerve, has been used as a simple measure of spinal motor neuron excitability. In ALS, an H reflex can sometimes be elicited from muscles, such as intrinsic hand muscles, in which it is difficult to demonstrate this response in healthy subjects.23 This finding suggests that there is segmental hyperexcitability even though a tendon tap cannot itself visibly excite a reflex. Presynaptic inhibition of the Group Ia synapse on motor neurons has been found to be significantly reduced in ALS,24 a finding not explained by motor neuron dropout in the disease but likely to be due to loss of Renshaw cells. Another possible reason for difficulty eliciting a tendon reflex in ALS is that changes in distal axon calibre may cause dispersion of the efferent volley, resulting in a non-synchronous muscle response. However, this potential functional abnormality should also make detection of the H response difficult.

Spasticity in ALS

Spasticity is often undetectable or absent in weak and wasted muscles in ALS.5 In primary lateral sclerosis, a disorder in which LMN signs are absent and the spinal motoneurons are intact, spasticity, often with rigidity, is a prominent clinical abnormality. Spasticity is a more complex response to UMN dysfunction than increased tendon reflex activity.10 Classically, spasticity consists of a velocity-dependent increase in muscle tone, associated with increased stretch responsiveness, and the clasp-knife phenomenon, representing autologous inhibition occurring in response to the imposed stretch.10 25 Spasticity is dependent on increased excitatory drive, and reduced inhibitory projections, on spinal segments. The dominant physiological abnormalities in spasticity have been summarised by DeSeilligny and Burke10 as decreased post-activation depression, facilitated Group I and Group II excitation and decreased recurrent Renshaw cell inhibition (see table 12.1 page 578 of Pierrot-Deseilligny and Burke). A number of ill-defined additional mechanisms are also operative, including increased α motor neuronal tone (‘α rigidity’), increased stiffness of muscles in the spastic limb,15 and increased non-corticospinal extrapyramidal discharge, leading to the phenomenon of ‘spastic rigidity’. However, classical spasticity differs from spastic rigidity and also from decerebrate rigidity, with the clasp-knife phenomenon,24 representing switching off the stretch-induced motor neuronal discharge by autogenetic inhibition. In ALS, the spinal motor neuron and interneuron wiring has been disrupted by the degenerative process, so that increased tone in this disease is variable and often atypical of the classical, velocity-dependent, stretch-sensitive, increased tone of spasticity as emphasised by the Sherrington school of physiology.25 There is no information on the relative pace of degenerative change in the α and γ motor neurons, or in internuncial neurons or Renshaw cells in the cord in ALS, but these factors will influence the output of the cord, including the development of spasticity.

The Babinski response in ALS

Finally, the plantar response, a reflex response to firm or even noxious cutaneous stimulation of the lateral border of the sole of the foot, is often unobtainable in ALS. The plantar response, of course, is not a stretch reflex, but a polysynaptic cutaneo-motor response. Lance26 has pointed out that the Babinski sign requires damage to the inhibitory cortico-reticulospinal pathway, located closely applied to the CST in its descending course. This leads to release of the segmental spinal motor system from descending inhibition, and therefore to release of flexor withdrawal reflexes, of which the Babinski response is a major example. Disorganisation of the α and γ motor neuron and interneuron connexions in the cord will affect this response. Although any relatively focal weakness from LMN lesion in ALS may affect the force vectors acting across the toes and ankle, unless very severe, this will not affect the recruitment pattern of muscles intrinsic to the flexor withdrawal reflex, and therefore the plantar response.


Signs of the UMN syndrome are difficult to elicit in ALS because the physiological basis for their release is itself disrupted by the degenerative process involving motor neurons of all classes in the anterior horn of the spinal cord. Direct assessment of the motor cortex or corticospinal pathways by transcortical motor stimulation,27 or by MRI, is probably the likely solution to this clinical problem. The widespread destruction of small motor neurons and their connexions in the anterior horns in ALS has been insufficiently recognised as a factor leading to the relatively partial development of classical UMN signs in the disease.


View Abstract


  • Competing interests None.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.