Stumbling reactions in man: significance of proprioceptive and pre-programmed mechanisms (Fragment)

BY V. DIETZ, J. QUINTERN AND M. SILLEM. From the Department of Clinical Neurology and Neurophysiology, University of Freiburg, Freiburg, F.R.G.


  1. Electromyogram (e.m.g.) responses of the leg musculature and the corresponding joint movements were studied following a perturbation of the limb during walking on a treadmill, produced by a randomly timed treadmill acceleration impulse, either predictable, or unpredictable in its amplitude and rate of acceleration.
    2. The rate of rise of ipsilateral gastrocnemius e.m.g. response following a perturbation was dependent on the rate of treadmill acceleration. For a given acceleration rate the amplitude of the e.m.g. response and the timing of its peak was dependent on the amplitude of the impulse and the rate of rise of the gastrocnemius response was the same for impulses of both small and large amplitude. The onset latency was shorter (65 ms) for high accelerations and longer (85 ms) for lower ones.
    3. The amplitude of the ipsilateral biceps femoris response was much smaller than the gastrocnemius response but was larger following unpredictable than predictable impulses. 4. The initial gastrocnemius response was followed by a tibialis anterior activation associated with a gastrocnemius depression and sometimes with a second, weak gastrocnemius activation. The gastrocnemius depression ended within a fixed time range relative to the onset of the response. The tibialis anterior activation was most pronounced when unpredictable impulses with high acceleration but a small amplitude were induced. 5. It is concluded that generation of the first gastrocnemius response is obviously under continuous control by muscle proprioceptive information and can be best described in terms of a stretch reflex response. It is suggested that, on the evidence of the diphasic or triphasic e.m.g. pattern, a close interaction occurs between a central programme and muscle proprioceptive input in order to generate the appropriate e.m.g. pattern.
    6. On the basis of earlier work (Berger, Dietz & Quintern, 1984a) and on the present results it is suggested that the e.m.g. responses may be mediated mainly by muscle proprioceptive input from group II afferents. This input is modulated and processed by spinal interneuronal circuits, closely connected with spinal locomotor centres. The mode of processing depends on various factors, such as the predictability of the nature of the impulse.


Perturbations of human gait unconsciously evoke complex and purposeful responses in leg, trunk and arm muscles in order to compensate for body imbalance and to prepare for any possible fall by extension of the arms. These reactions adapt quickly (within few trials) to a given mode of perturbation (Nashner, 1976; Nashner, Woollacott & Tuma, 1979; Forssberg & Nashner, 1982; Quintern, Berger & Dietz, 1985) and therefore cannot be explained on the basis of a simple stimulus-response relationship. Rather, they represent the result of more complex processing of afferent information. It is not, therefore, surprising that, for instance, segmental stretch reflex responses are inhibited during gait (Dietz, Quintern & Berger, 1984). Electromyogram (e.m.g.) responses in the leg following gait perturbations are qualitatively different from those described in arm muscles following a displacement (Berardelli, Hallett, Kaufmann, Fome, Berenberg & Simon, 1982; Berardelli, Sabra, Hallett, Berenberg & Simon, 1983; Berger et al. 1984a). It is therefore not justifiable to use the terminology coined for the latter (M1-M3; introduced by Lee & Tatton, 1975) to describe responses in the former. Although the latencies of the responses were shown to depend on the position when slow perturbations occurred during stance (Diener, Bootz, Dichgans & Bruzek, 1983), functionally appropriate e.m.g. responses in the leg generally appear with a short latency (60-75 ms) on both sides when rapid perturbations are induced during stance and gait (Berger et al. 1984a; Dietz et al. 1984). This makes it unlikely that the first part of the response, at least, is mediated by a transcortical pathway.