The early history of the knee-jerk reflex in neurology

ABSTRACT

Medical interest in the knee-jerk reflex began in about 1875 with simultaneous and independent publications by Wilhelm Heinrich Erb (1840–1921) and Carl Friedrich Otto Westphal (1833–1890) contending that the knee jerk was absent (and the ankle clonus was present) in all clear cases of locomotor ataxia (tabes dorsalis). Physicians in the medical communities of Europe, Great Britain, and North America responded with case and large group studies that tested this contention. These studies revealed the usefulness of the knee jerk and other myotatic reflexes, but also unexpected characteristics. The knee jerk, apparently so simple, proved to be a complex phenomenon depending the strength of the strike on the patella, induced muscle tension, and inhibition from the brain. Was it a reflex with afferent and efferent nerves and an intervening process in the spinal cord, or was it a local phenomenon confined to the muscle itself? Experimental studies directed at the reflex issue investigated latencies from patella strike to leg extension or muscle contraction and compared them with latencies from direct muscle strikes and theoretical calculations based on reflex components. Such studies were unable to resolve the reflex issue during the nineteenth century. The physicians were shown to be limited, like all scientific explorers of the unknown, by their knowledge, methodology, and technology.

KEYWORDS:

• Experimental
• knee jerk
• locomotor ataxia
• medical research
• myotatic reflex
• nineteenth century
• patella ligament
• reflex
• tabes dorsalis

Acknowledgments

I wish to thank Drs. Ed Fine and Carol Diakow for their helpful suggestions on early drafts of this article.

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

¹ The absence of this reflex can occur in healthy individuals; however, more likely, when in the context of other signs its absence can be due to a variety of causes, including several reasons probably appreciated and so named by the end of the nineteenth century, such as a crush injury of the spinal cord destroying motor neurons supplying the femoral nerve; severe myopathy, which may reduce or eliminate the reflex; radiculopathy of L2-L4 roots; and peripheral neuropathy. Of course, there are other reasons possibly not named as such or not known during the time considered in this essay, such as lumbar plexopathym dorsal column disease of the spinal cord, and ganglionopathy affects due to input of Ia fibers into the spinal cord (personal communication, anonymous).

² The presence of the Achilles clonus is indicative of impairment, whereas its absence is normal. The Achilles tendon was forced to remain flexed consequently the rhythmic clonus of the foot. The Achilles clonus phenomenon must be distinguished from the Achilles jerk or ankle jerk, which results from a tap on the Achilles tendon when the foot is dorsal flexed. It is present or absent in the same manner as the knee jerk. The plantar reflex is also associated with the foot. It is elicited by stimulation of the sole of the foot. The response is flexion or extension of the toes. Flexion is the normal adult response, whereas extension is pathological, but normal in infants. Extension of the toes in this context is called the Babinski response or Babinski sign (Wolff 1930).

³ Gowers’s patients were differentiated only as 150 with no known paralytic condition of the lower extremities, with the majority of these 150 identified as epileptics and some “in perfectly normal health.” The diagnoses of the other 150 were not mentioned. He apparently published data in this form because of the paucity of information about the knee jerk in the literature. Nonetheless, its value is diluted by the group’s mix of unknown etiologies.

⁴ In her article about the Clinic for Nervous Diseases of the Moscow University, Vein commented that in 1892 A. Y. Kozhevnikov stated that syphilis was the main cause of tabes dorsalis.

⁵ To what extent were the contradictory results due to unknown aspects of the knee jerk? Does the position of the leg make a difference? Was the variability due to how hard the tendon was struck or where it was struck? Did the tendon need to be struck at all? Did the pathology of the patient have an effect? What effect did superior parts of the nervous system have? What were the parameters of the reinforcing act?

⁶ “… während ich auf seine Patellarsehne klopfe, fordere ich es auf, die gebeugten Finger der rechten und linken Hand in einander zu hängen und sie bei nach vorne ausgestreckten (horizontal) Armen so stark als möglich auseinander zu ziehen.” (“… while I knock on his patellar tendon, I ask him to hang the bent fingers of the right and left hands together and pulling them apart as much as possible with arms stretched forward [horizontally])” (Jendrássik 1885, 413).

⁷ Hundreds of measurements were thought necessary for reliability. Outside of consistent treatments and replications, he had no other way of demonstrating reliability.

⁸ According to Fearing (1928), this study was particularly important to future researchers. It was referenced in physiological texts by Howell, Luciani, Ladd and Woodworth, and Schaifer. Because of its importance, he analyzed it and found methodological problems.

⁹ The introduction of the bell that preceded the response is startling in light of current knowledge about classical conditioning. However, the reason that Bowditch and Warren introduced the bell was to standardize the onset of the reinforcement. At the sound of the bell, the subject was to clinch his right fist around an “electric key” that closed an electrical circuit and began a reaction time measure. The response to the bell insured that all aspects of the reaction time was included for “a reaction time always includes a centripetal, a central and a centrifugal portion, and as it was impossible to say in advance which portion of this nervous process would have the greatest influence upon the knee-jerk” (Bowditch and Warren 1890, 36–37).

¹⁰ The modern latency from the stretch of the patellar tendon to the beginning of the contraction is about 18 ms. From the modern point of view, their theoretical component analyses were way out of line.

¹¹ The string galvanometer was available at the time, but none of the latency measures were acquired with that device during the period under discussion (Sternberg 1893).

¹² This would give a resolution of 10 ms +– 5 ms. Sternberg (1893) discussed a kymograph used by Brissaud in 1880 that had a resolution of 2 ms +– 1 ms.

¹³ “Helmholtz’s original experiments gave for the nerve transmission-rapidity 30 to 90 m. per sec. according to temperature. Bloch had stated that the rate might reach over 150 m. per sec” (Waller 1880). Based on the 1.5 meters of nerve fiber, these would make the range of transmission times between 10 ms and 50 ms. The present times for 1a fibers is between 80 and 120 m/s. For 1.5 m of nerve fiber, transmission time would be between 13 ms and 19 ms.

¹⁴ For chemical synapses, according to modern reckoning, there is a delay (usually, approximately .5–1 ms in duration) between the initiation of an action potential in the presynaptic terminal and a potential change in the postsynaptic cell.

¹⁵ Possibly off by a factor of 10 according to modern measures.

Additional information

Funding

The authors reported there is no funding associated with the work featured in this article.