Antagonists’ alternation during voluntary oscillations of the extremities adapts to the mechanical context. Experimental evidences and a neural control model

Many common life gestures require the ability to combine voluntary movements of different limb segments into a variety of co-ordinated patterns. When two limb segments are moved together, non-mechanical constraints act to facilitate certain associations and to hinder, or even impede others. The effect of such constraints is especially manifest when coupling isorhythmic oscillations. For example, it is quite simple to couple flexion of the hand prone with plantarflexion of the foot, while it is apparently more difficult to couple hand extension with foot plantarflexion. A similar behaviour is also shown by many other associations of limb segments. When studying coupled oscillations of the ipsilateral hand and foot, Baldissera et al. (1991, 2000; Baldissera and Cavallari 2001) measured the inter-limb phase-relations, both between hand and foot movements and, respectively, between EMG onset in Extensor Carpi Radialis (ECR) and Tibialis Anterior (TA). Moreover, they described the phase-relation intrinsic to each of the two extremities, expressed as the frequency-dependent (phase) delay between the onset of EMG activity and the onset of the related movement. In the hand, such relation could be well fitted by a pendulum model, while this was not possible for the foot oscillations, since at low-frequency (<1.5 Hz) onset of TA EMG paradoxically phase-lagged the onset of dorsiflexion, suggesting that the initial part of the movement was sustained by the recoil of elastic structures that were stretched during plantarflexion. Since the recoil should return the moving segment to its equilibrium position, i.e. that position in which the limb rests when muscles are fully relaxed, it may be argued that TA contraction was only needed to move the foot away from the equilibrium. As a consequence, the phase lag should disappear if the EMG onset were referred to the foot passive equilibrium position rather than to the movement onset. This hypothesis was tested in 10 subjects who voluntarily oscillated their right foot at various frequencies (0.2 to 3Hz) over 3 angular ranges: a central-range (foot crossing the equilibrium symmetrically), a high-range (whole excursion above equilibrium) and a low-range (whole excursion below equilibrium). In the central-range, phase-relations were measured between the crossing of equilibrium position and the onset of the TA EMG during dorsiflexion or the onset of Soleus (Sol) EMG during plantarflexion. In both cases, the phase-curves started around zero, without showing any paradoxical lag of EMG on movement. Phase-curves with similar features were also obtained in the high- and low-ranges (no crossing of equilibrium) by correlating the onset of the EMG burst to the onset of the related movement. The patterns of muscle activation recorded in all these conditions may be conceived as the result of one single motor output, which is split and distributed to the antagonists as soon as the segment crosses its equilibrium position. However, the passive equilibrium position may vary following the application of external loads or changes in limb orientation. In the hand, for instance, the equilibrium is reached in one single position when the hand is prone (flexion-extension vertical) but it covers an angular range when the hand is semi-prone, i.e., midway between prone and supine (flexion-extension horizontal). The analysis performed on foot oscillations was thus extended to the hand to determine how antagonists’ alternation is regulated in the presence of an equilibrium range. Activity distribution between ECR and Flexor Carpi Radialis (FCR) during rhythmic flexion-extension of the hand was analysed in three different mechanical conditions (5 subjects). In the first condition (hand prone, flexion-extension in a vertical parasagittal plane) the hand passive equilibrium position was ~50° in flexion. During hand oscillations FCR and ECR were alternatively recruited to move the hand symmetrically away from the equilibrium and de-recruited to allow conservative forces to restore the equilibrium. Switching between antagonists occurred at the centre of the oscillation (equilibrium crossing), just like it happened for the foot. In the second condition (hand semi-prone, flexion-extension in a horizontal transversal plane) the hand equilibrium was attained over an angle of about 26°. When the hand was oscillated symmetrically around this equilibrium range, each muscle was recruited when the hand entered the equilibrium range, i.e. in advance of the oscillation centre. Both vertical and horizontal oscillations were also performed all externally to the equilibrium position or range: in these cases only one muscle was recruited over the entire cycle, the EMG burst starting at the onset of the related movement. In the third condition (same orientation as condition 2) the application of a frictional load expanded the equilibrium range to encompass the entire hand oscillation. Now concentric muscle contraction was needed throughout each phase of the movement and switching between antagonists occurred at the movement reversal, i.e. ~90° in advance of the oscillation centre. Altogether, these findings show that during both hand and foot oscillations contractile force starts developing when an intrinsic or external resistance has to be overcome in order to continue the movement. This behaviour indicates that the nervous system monitors the mechanical characteristics of the moving limb and consequently adapts to them the pattern of antagonists’ activation. To account for such a control, a neural network is proposed that compares the afferent information about joint position with a position central command, thus detecting the position error caused by the forces that resist to movement. From the sign and amplitude of the error signal, the network determines the direction (agonist vs antagonist) and the amount of motor activation required. The control capability of the network was tested by connecting it to a realistic model of the limb, which accounted for the mechanical properties of the segment and, when present, of the load. When simulating each of the mechanical conditions in which the hand or the foot were oscillated, the proposed neural network reproduced the patterns of antagonists’ alternation observed in the real experiments.