Altered accelerator pedal control in a driving simulator in people with diabetic peripheral neuropathy.

AIM: To investigate whether the sensory-motor impairment attributable to diabetic peripheral neuropathy would affect control of the accelerator pedal during a driving simulator task. METHODS: A total of 32 active drivers, 11 with diabetic peripheral neuropathy (mean ± sd age 67±5.0 years), 10 with diabetes but no neuropathy (diabetes group; mean ± sd age 62±10 years), and 11 healthy individuals without diabetes (healthy group; mean ± sd age 60±11 years), undertook a test on a dynamometer to assess ankle plantar flexor muscle strength and ankle joint proprioception function of the right leg, in addition to a driving simulator task. The following variables were measured: maximal ankle plantar flexor muscle strength; speed of strength generation (Nm/s); and ankle joint proprioception (ankle repositioning error, degrees). In the driving simulator task, driving speed (mph), accelerator pedal signal (degrees) and the duration of specific 'loss-of-control events' (s) were measured during two drives (Drive 1, Drive 2). RESULTS: Participants with diabetic peripheral neuropathy had a lower speed of strength generation (P<0.001), lower maximal ankle plantar flexor muscle strength (P<0.001) and impaired ankle proprioception (P=0.034) compared to healthy participants. The diabetic peripheral neuropathy group drove more slowly compared with the healthy group (Drive 1 P=0.048; Drive 2 P=0.042) and showed marked differences in the use of the accelerator pedal compared to both the diabetes group (P=0.010) and the healthy group (P=0.002). Participants with diabetic peripheral neuropathy had the longest duration of loss-of-control events, but after one drive, this was greatly reduced (P=0.023). CONCLUSIONS: Muscle function, ankle proprioception and accelerator pedal control are all affected in people with diabetic peripheral neuropathy, adversely influencing driving performance, but potential for improvement with targeted practice remains possible.

People with diabetic peripheral neuropathy display a decreased stepping accuracy during walking: potential implications for risk of tripping.

AIM: To examine the stepping accuracy of people with diabetes and diabetic peripheral neuropathy. METHODS: Fourteen patients with diabetic peripheral neuropathy (DPN), 12 patients with diabetes but no neuropathy (D) and 10 healthy non-diabetic control participants (C). Accuracy of stepping was measured whilst the participants walked along a walkway consisting of 18 stepping targets. Preliminary data on visual gaze characteristics were also captured in a subset of participants (diabetic peripheral neuropathy group: n = 4; diabetes-alone group: n = 4; and control group: n = 4) during the same task. RESULTS: Patients in the diabetic peripheral neuropathy group, and patients in the diabetes-alone group were significantly less accurate at stepping on targets than were control subjects (P < 0.05). Preliminary visual gaze analysis identified that patients diabetic peripheral neuropathy were slower to look between targets, resulting in less time being spent looking at a target before foot-target contact. CONCLUSIONS: Impaired motor control is theorized to be a major factor underlying the changes in stepping accuracy, and potentially altered visual gaze behaviour may also play a role. Reduced stepping accuracy may indicate a decreased ability to control the placement of the lower limbs, leading to patients with neuropathy potentially being less able to avoid observed obstacles during walking.

Eye-steering coordination in natural driving.

When driving along a winding road, eye movements and steering are tightly linked; the driver looks across to the inside kerb of an approaching bend some time before turning the steering wheel. With the eyes leading, the oculomotor controller assists the neural centres controlling steering; prevention of any eye movements correlated with steering impairs driving, so the coordination is crucial for safety. A key question is therefore what are the limits of acceptable variation in timing and degree of coordination. Over a period of continuous driving on the open road, how much does the relative timing and degree of coordination between eye and steering movements vary? A related question is how brief a period of driving will suffice to measure these coordination parameters. Drivers' eye movements and steering were measured over different time periods ranging from 15 s to 6 min epochs of natural driving along a winding country road to establish the variability in coordination and the minimum time period required to characterise it. We show here that brief periods of driving, 30 s or less, are inadequate for describing eye-steering coordination. But a minute of driving yields an accurate description much of the time; and 2 min is sufficient both to accurately describe this relationship and to show that it is highly consistent for a given individual, and for different people driving the same route.

Eye movements coordinated with steering benefit performance even when vision is denied.

When driving along a winding road, eye movements and steering are tightly linked. When approaching a bend, the driver looks across to the inside kerb (the tangent point) some time before turning the steering wheel. All drivers we have tested show this optimal coordination, without which driving is impaired. An intriguing question is how much of the benefit for steering arises just from moving the eyes in this coordinated way (ahead of steering and in the same direction), and how much from the visual information that the eyes move to acquire, in this instance the foveated tangent point. This can be answered by dissociating the two, by reducing visibility of the road ahead (and crucially of the tangent point) to a level at which drivers might or might not choose to move their eyes but, if they do, will not gain the information they seek. Twenty subjects repeatedly drove a simulated stage of the World Rally Championship. With full visibility, they exhibited the usual coordination of eye movements and steering. Subsequently, visibility was reduced on the left hand side. Drivers who persisted in making eye movements coordinated with steering to the left, despite the fact that they could no longer see the tangent point on that side, performed better than drivers who under the identical conditions did not look to the left. This confirms that the making of coordinated eye movements itself benefits steering, even when the eye movements do not yield the visual information sought.

Prevention of coordinated eye movements and steering impairs driving performance.

When approaching a bend in the road, a driver looks across to the inside kerb before turning the steering wheel. Eye movements and steering are tightly linked, with the eyes leading, which means that the oculomotor controller can assist the neural centres controlling steering. This optimum coordination is observed for all drivers; but despite being the preferred solution to the motor-control problem of successfully steering along a winding road, the question remains as to how crucial such coordinated eye and steering movements are for driving performance. Twenty subjects repeatedly drove a simulated stage of the World Rally Championship, aiming to complete the course in the fastest possible time. For the first six repetitions they used the usual coordination of eye movements and steering; for drives 7--12 they were instructed to fixate on a small spot in the centre of the screen (centre gaze). Prevention of coordination in this way impaired their performance (drives 6 and 7 compared), dramatically increasing their time taken to complete the course, equivalent to slipping 19 places down the leader board in the actual rally stage. This indicates that the usual pattern of eye movements correlated with steering is crucial for driving performance. Further experiments are suggested to reveal whether any attentional demand associated with keeping the eyes still contributes to the loss in performance.

The lateral cerebellum and visuomotor control.

The lateral cerebellum receives an abundance of visual input providing the link between visual and motor control centers. In this review we discuss experiments designed to increase our understanding of how visual inputs to the cerebellum are arranged in relation to the zonal organization of the cerebellar cortex, and how visual inputs are utilized to assist in the regulation of a visually guided movement. On the basis of anatomical and physiological characteristics our findings indicate that the medial-most folium in crus I of the cat lateral cerebellum can be subdivided into at least three functionally distinct zones; from lateral to medial along the length of the folium these correspond to zones D(1), lateral C(3) and C(2). Each zone displays clear differences in olivo-cortico-nuclear connectivity and in the anesthetized animal zones D(1) and C(2) both receive powerful visual inputs relayed via the climbing fiber system. Complementary experiments in awake behaving cats found that Purkinje cells located in the D(1) and D(2) zones of crus I exhibit changes in simple spike discharge time locked to target motion during a visually guided reaching task. These changes were unaffected by temporary visual denial of the target, raising the possibility that internally generated feedforward visuomotor control mechanisms are operating, in which a predictive model of the target's motion has been constructed by the CNS.

Direct visualisation of gaze and hypometric saccades in cerebellar patients during visually guided stepping.

Four patients suffering from primary cerebellar degeneration and healthy matched controls undertook a test of functional mobility that demanded precise foot placement at each step. Vertical and horizontal eye movements were measured (using a head mounted eye tracking system) together with footfall patterns. Healthy subjects stepped accurately onto all targets and produced a clear pattern of saccadic eye movements, fixating each target in the sequence just prior to footlift. Still video frames, showing direction of gaze while walking, provide direct visual confirmation that these saccades serve to transfer gaze between successive targets in the walkway sequence. The planning of the saccade to the next target probably provides the locomotor control system with information useful for planning the corresponding (and shortly following) step. Cerebellar patients showed characteristic locomotor and oculomotor deficits. Dysmetric saccades to fixate footfall targets were seen in 39% of steps. Analysis confirms that these multi-saccadic eye movements include an initial hypometric saccade, which undershoots the target, followed by one or more additional saccades in the same direction. Direct visualisation of gaze at the end of a saccadic sequence confirms that these additional saccades are indeed corrective resulting in a foveal image of the footfall target.

Alcohol affects eye movements essential for visually guided stepping.

BACKGROUND: To understand how and why alcohol intoxication affects visually guided stepping, the eye movements and performance of 6 subjects (aged 22-35 years) were monitored as they progressed along a pathway of 18 irregularly placed stepping stones before and after consumption of an acute oral dose of alcohol. METHODS: Horizontal eye movements were measured with infrared reflectometry; footfall on or off target was monitored via copper fabric soles stuck to subjects' footwear. Breath alcohol concentration was monitored with an Alco-Sensor III breathalyzer. RESULTS: After alcohol loading, both locomotor and oculomotor deficits were evident. All subjects increased their step cycle duration-with prolonged stance, swing, and double support phases-and occasionally missed footfall targets. A large proportion of saccades made to fixate successive stepping stones were inaccurate and were accompanied by one or more corrective saccades. These problems with looking and stepping to footfall targets tended to occur together and were comparable to those seen previously in cerebellar patients undertaking the same task. CONCLUSIONS: The fact that healthy subjects acutely intoxicated by alcohol show symptoms of cerebellar dysfunction suggests that alcohol acutely and adversely affects the cerebellar contribution to performance of visually guided movements.

Rehearsal by eye movement improves visuomotor performance in cerebellar patients.

In order to assess the effect of rehearsal by eye movement alone on visuomotor performance, the eye movements and visually guided stepping of two cerebellar patients were monitored before and after a first and second batch of eye-movement rehearsals, in which patients made saccadic eye movements to the first 6 footfall targets (in a sequence of 18) whilst standing stationary at the start of the walkway. There was a marked improvement in oculomotor and locomotor performance following the second batch of eye-movement rehearsal. Both patients showed reduced occurrence of saccadic dysmetria, evident as a significant increase in the proportion of single to multi-saccadic eye movements (from 46 to 77% for DB and from 75 to 94% for TP). This was accompanied by increased regularity and accuracy of stepping in both patients, and decreased stance and double support phase durations (one patient only). Separate testing confirmed that these improvements in eye movements and stepping did not result from simple repetition of the task. This is the first demonstration of a technique--rehearsal by eye movement--that improves the visuomotor performance of cerebellar patients. It is compelling evidence for our proposal that during visually guided stepping the locomotor control system is dependent on assistance from the oculomotor control system.

Coordination of eye and leg movements during visually guided stepping.

In the present study, 2 related hypotheses were tested: first, that vision is used in a feedforward control mode during precision stepping onto visual targets and, second, that the oculomotor and locomotor control centers interact to produce coordinated eye and leg movements during that task. Participants' (N = 4) eye movements and step cycle transition events were monitored while they performed a task requiring precise foot placement at every step onto irregularly placed stepping stones under conditions in which the availability of visual information was either restricted or intermittently removed altogether. Accurate saccades, followed by accurate steps, to the next footfall target were almost always made even when the information had been invisible for as long as 500 ms. Despite delays in footlift caused by the temporary removal (and subsequent reinstatement) of visual information, the mean interval between the start of the eye movement and the start of the swing toward a target did not vary significantly (p >.05). In contrast, the mean interval between saccade onset away from a target and a foot landing on that target (stance onset) did vary significantly (p <.05) under the different experimental conditions. Those results support the stated hypotheses.

Evidence for interactive locomotor and oculomotor deficits in cerebellar patients during visually guided stepping.

Eight patients suffering from primary cerebellar degenerative diseases undertook a walkway task, demanding precise foot placement at each step, and a visual fixation task, requiring only eye movements. Step cycle and horizontal eye movements were recorded throughout the tasks and compared to those of healthy adults (including age- and sex-matched controls). Cerebellar patients displayed both locomotor and oculomotor deficits. Increases in duration of the stance, swing and double support phases of the step cycle were all shown to contribute to ataxic gait. Dysmetric saccades to fixate the footfall targets were seen more frequently in patients than in controls. These hypometric saccades were followed by one or more corrective saccades (patients: >45% accompanied by one or more corrective saccades; controls: <10% accompanied by a single corrective saccade). Similarities between the oculomotor deficits displayed by patients during the visual fixation task and when walking indicate that the latter are not merely a consequence of ataxic gait. The existence of several links between these locomotor and oculomotor deficits provides evidence for considerable interaction between the two control systems in the production of patterned eye and stepping movements. These results also suggest that the cerebellum plays an active role in the co-ordination of visually guided eye and limb movements during visually guided stepping.

Central regulation of motor cortex neuronal responses to forelimb nerve inputs during precision walking in the cat.

1. The responses of neurones in forelimb motor cortex to impulse volleys evoked by single pulse electrical stimulation (at 1.5 or 2 times the threshold for most excitable nerve fibres) of the superficial radial (SR) and ulnar (UL) nerves of the contralateral forelimb were studied in awake cats both resting quietly and walking on a horizontal ladder. Nerve volley amplitude was monitored by recording the compound action potential elicited by the stimulus. 2. In the resting animal 34/82 (41%) cells yielded statistically significant responses to SR stimulation, and 20/72 (28%) responded to UL stimulation. Some responses were confined to or began with an increase in firing probability ('excitatory' responses) and others with a decrease in firing ('inhibitory' responses), typically including a brief interruption of the spike train (zero rate). Cells responding to both nerves usually yielded responses similar in type. Most (78%) response onset latencies were less than 30 ms. Responses involved the addition or subtraction of from 3.4 to 0.1 impulses stimulus-1 (most <1 impulse stimulus-1). The distribution of response sizes was continuous down to the smallest values, i.e. there was no 'gap' which would represent a clear separation into 'responsive' and 'unresponsive' categories. Responses were commonest in the lateral part of the pericruciate cortex, and commoner among pyramidal tract neurones (PTNs) than non-PTNs. 3. During ladder walking most cells generated a rhythmic step-related discharge; in assessing the size of responses to nerve stimulation (20 studied, from 13 cells) this activity was first subtracted. Response onset latencies (90% <30 ms) and durations showed little or no change. Although most cells were overall more active than during rest both 'excitatory' and 'inhibitory' responses in both PTNs and non-PTNs were often markedly reduced in large parts of the step cycle; over some (usually brief) parts responses approached or exceeded their size during rest, i.e. response size was step phase dependent. Such variations occurred without parallel change in the nerve compound action potential, nor were they correlated with the level of background firing at the time that the response was evoked. When responses to both nerves were studied in the same neurone they differed in their patterns of phase dependence. 4. The findings are interpreted as evidence for central mechanisms that, during 'skilled', cortically controlled walking, powerfully regulate the excitability of the somatic afferent paths from forelimb mechanoreceptors (including low threshold cutaneous receptors) to motor cortex. Retention (or enhancement) of responsiveness often occurred (especially for ulnar nerve) around footfall, perhaps reflecting a behavioural requirement for sensory input signalling the quality of the contact established with the restricted surface available for support.

Rhythmic neuronal activity in the lateral cerebellum of the cat during visually guided stepping.

1. The discharge patterns of 117 lateral cerebellar neurones were studied in cats during visually guided stepping on a horizontal circular ladder. Ninety per cent of both nuclear cells (53/59) and Purkinje cells (53/58) showed step-related rhythmic modulations of their discharge frequency (one or more periods of 'raised activity' per step cycle of the ipsilateral forelimb). 2. For 31% of nuclear cells (18/59) and 34% of Purkinje cells (20/58) the difference between the highest and lowest discharge rates in different parts of the step cycle was > 50 impulses s-1. 3. Individual neurones differed widely in the phasing of their discharges relative to the step cycle. Nevertheless, for both Purkinje cells and nuclear cells population activity was significantly greater in swing than in stance; the difference was more marked for the nuclear population. 4. Some cells exhibited both step-related rhythmicity and visual responsiveness (28 of 67 tested, 42%), whilst others were rhythmically active during locomotion and increased their discharge rate ahead of saccadic eye movements (11 of 54 tested, 20%). The rhythmicity of cells that were visually responsive was typical of the rhythmicity seen in the whole locomotor-related population. The step-related rhythmicity of cells that also discharged in relation to saccades was generally below average strength compared with the cortical and nuclear populations as a whole. 5. The possibility is discussed that the rhythmicity of dentate neurones acts as a powerful source of excitatory locomotor drive to motor cortex, and may thereby contribute to establishing the step-related rhythmicity of motor cortical (including pyramidal tract) neurones. More generally, the activity patterns of lateral cerebellar neurones provide for a role in the production of visually guided, co-ordinated eye and body movements.

Neuronal activity in the lateral cerebellum of the cat related to visual stimuli at rest, visually guided step modification, and saccadic eye movements.

1. The discharge patterns of 166 lateral cerebellar neurones were studied in cats at rest and during visually guided stepping on a horizontal circular ladder. A hundred and twelve cells were tested against one or both of two visual stimuli: a brief full-field flash of light delivered during eating or rest, and a rung which moved up as the cat approached. Forty-five cells (40%) gave a short latency response to one or both of these stimuli. These visually responsive neurones were found in hemispheral cortex (rather than paravermal) and the lateral cerebellar nucleus (rather than nucleus interpositus). 2. Thirty-seven cells (of 103 tested, 36%) responded to flash. The cortical visual response (mean onset latency 38 ms) was usually an increase in Purkinje cell discharge rate, of around 50 impulses s-1 and representing 1 or 2 additional spikes per trial (1.6 on average). The nuclear response to flash (mean onset latency 27 ms) was usually an increased discharge rate which was shorter lived and converted rapidly to a depression of discharge or return to control levels, so that there were on average only an additional 0.6 spikes per trial. A straightforward explanation of the difference between the cortical and nuclear response would be that the increased inhibitory Purkinje cell output cuts short the nuclear response. 3. A higher proportion of cells responded to rung movement, sixteen of twenty-five tested (64%). Again most responded with increased discharge, which had longer latency than the flash response (first change in dentate output ca 60 ms after start of movement) and longer duration. Peak frequency changes were twice the size of those in response to flash, at 100 impulses s-1 on average and additional spikes per trial were correspondingly 3-4 times higher. Both cortical and nuclear responses were context dependent, being larger when the rung moved when the cat was closer than further away. 4. A quarter of cells (20 of 84 tested, 24%) modulated their activity in advance of saccades, increasing their discharge rate. Four-fifths of these were non-reciprocally directionally selective. Saccade-related neurones were usually susceptible to other influences, i.e. their activity was not wholly explicable in terms of saccade parameters. 5. Substantial numbers of visually responsive neurones also discharged in relation to stepping movements while other visually responsive neurones discharged in advance of saccadic eye movements. And more than half the cells tested were active in relation both to eye movements and to stepping movements. These combinations of properties qualify even individual cerebellar neurones to participate in the co-ordination of visually guided eye and limb movements.

A method for automatic identification of saccades from eye movement recordings.

We describe a technique for reliable and rapid automatic identification of saccades in eye movement records. The signal processing that we describe will be useful to anyone wanting to analyse large numbers (thousands) of eye movements. We describe a transform that is derived from the differentiated eye movement record, and which is related to a transform previously used to automate analysis of EMG recordings.

Visually guided stepping under conditions of step cycle-related denial of visual information.

We recently reported that subjects performing a task that requires visual guidance of each step onto irregularly placed "stepping stones" usually fixate the next target of footfall just before they lift the foot to be repositioned, i.e. towards the end of that limb's stance phase. When negotiating the same walkway without ambient lighting, and with each stone's location indicated by a central light spot (LED), stepping and eye movements were unchanged. Under conditions of intermittent visual denial, in which all LEDs (the only visual cues) were temporarily extinguished at irregular intervals, temporal changes in the normal stepping pattern were sometimes observed, but stepping was not always affected. The primary effect of visual denial was on the leg that was in stance (foot in place on a stepping stone) at the moment of LED extinction, rather than on the leg that was in swing, and was an increase in stance duration, suggesting an effect on planning during this stance of the next swing towards the next target rather than on execution of the ongoing swing of the other leg. Subjects rarely failed to step onto the targets. Prolongations of stance under visual denial lasting 400 or 500 ms were less than 200 ms, much less than the duration of denial; subjects did not simply wait for the footfall target to reappear. There was no effect for denial lasting 300 ms; subjects performed as well as with a constantly visible target. Under 400 and 500 ms denial, there was no effect when the targets disappeared in the first 100 ms of stance (of the foot to be repositioned); stance durations were indistinguishable from control. This suggests that there is no crucial visuomotor processing by the control system(s) for eye and limb guidance until the target reappeared near the usual end of stance, when feedforward planning of the next saccade and/or swing to a target reaches a crucial stage, and is affected by intrusion of the period of visual denial. With longer (800 ms) denial there was an effect regardless of when in stance it began. A smaller effect of 800 ms denial sometimes visible in swing duration is attributable to interlimb coordination. Accurate saccades, followed by accurate steps, to the next target are almost always made, even when the target is invisible. Our results demonstrate that uninterrupted on-line visual information is not necessary for accurate stepping even when (as here) each step requires visual guidance. Also, since stance prolongations did not always result, and they were always much shorter than the periods of denial, we conclude that the visuomotor control mechanism(s) are robust in the face of substantial denial of all visual information including normally preferred inputs (foveal or peripheral images) at the normally preferred times. The fact that a saccade is still made to an invisible target location implies that this is useful in itself, since it does not result in a visible foveal image. We propose that skilled, visually guided stepping onto irregularly placed targets is executed under predominantly feedforward visuomotor control mechanisms, and suggest that the ability to function effectively in this way is dependent upon the integrity of the lateral cerebellum.

Role of the cerebellum and motor cortex in the regulation of visually controlled locomotion.

An account is given of the current state of knowledge of the contributions of the cerebellum and the forelimb motor cortex (MC) to the neural control of walking movements in the cat. The main emphasis is on information obtained by recording from single MC and cerebellar neurones in chronically instrumented cats engaged in walking on the rungs of a horizontal ladder, a form of locomotion that is heavily dependent on visual input and for which the integrity of MC is essential. Evidence from the authors' laboratory and from other studies is presented which establishes that MC neurones, including pyramidal tract neurones, show higher levels of activity during ladder walking than during overground walking (i.e., when less constraint exists over the locus of footfall) and that this increase is greatest in late swing-early stance in the contralateral forelimb, consistent with one role of MC being to help determine the locus of footfall. However, many MC neurones develop peak activity at other times in the step cycle, and a comparison with recordings during treadmill walking suggests MC may also help regulate stance duration when walking speed is an important performance variable. Recordings from Purkinje cells and cerebellar nuclear neurones show that during ladder walking step-related activity is widespread in the vermal, paravermal, and crural regions of cortex and in the interposed and dentate nuclei. Nuclear cell activity is so timed that it could be contributing to producing the locomotor rhythms evident in MC cells, although this is not yet proven. Results are also presented and discussed relating to MC and cerebellar neuronal responses that occur when a step onto an unstable rung results in an unexpected external perturbation of the forelimb step cycle. MC responses begin with onset latency as short as 20 ms so that MC may assist spinal reflex mechanisms to produce a post hoc compensatory change in motor output. However, work in progress suggests that corresponding responses in paravermal cerebellum are weak and infrequent, so provisionally it seems that the MC responses are initiated via pathways that do not pass through the cerebellum. By contrast, current work involving a paradigm in which a ladder rung is motor driven to a new position as the animal approaches (thereby providing a visual cue that an adaptive change in gait will soon be required) is revealing in lateral cerebellar neurones, including dentate neurones, changes in discharge that are time locked to the execution of an adapted pace. In addition, there are prominent earlier responses, which begin at short latency after the onset of rung movement. These apparently visual responses have characteristics that encourage the speculation that they may represent a cerebellar signal that "primes for action" other more directly motor regions of the central nervous system.

Human Eye Movements During Visually Guided Stepping.

Visually guided locomotion was studied in an experiment in which human subjects (N = 8) had to accurately negotiate a series of irregularly spaced stepping-stones while infrared reflectometry and electrooculography were used to continuously record their eye movements. On average, 68% of saccades made toward the next target of footfall had been completed (visual target capture had occurred) while the foot to be positioned was still on the ground; the remainder were completed in the first 300 ms of the swing phase. The subjects' gaze remained fixed on a target, on average, until 51 ms after making contact with it, with little variation. A greater amount of variation was seen in the timing of trailing footlift relative to visual target capture. Assuming that subjects sampled the visual cues as and when they were required, visual information appeared most useful when the foot to be positioned was still on the ground.

Changes in the discharge patterns of cat motor cortex neurones during unexpected perturbations of on-going locomotion.

1. The impulse activity of single neurones in the forelimb part of the motor cortex was recorded extracellularly in unrestrained cats during self-paced locomotion on a horizontal circular ladder. 2. Fifty-one cells (forty-nine of which discharged rhythmically in time with the step cycle) were recorded during encounters with a number of rungs that could be locked firmly in position or, alternatively, held in position by weak springs so that when stepped on they unexpectedly descended (under the weight of the animal) from 1 to 5 cm before contacting a mechanical stop. 3. In eleven cells (22%) including four fast-axon pyramidal tract neurones (PTNs), an increase in discharge occurred when the contralateral forelimb descended unexpectedly. Onset latency relative to the start of rung movement ranged from ca 20 to ca 100 ms. In eight cells latency was such that most of the response preceded contact of the rung with the stop; averaged over a number of trials the altered discharge in five of these cells (including two PTNs) represented an accurate profile of the averaged velocity of rung (and foot) descent. The three remaining cells appeared to be responding largely to the cessation of rung movement. 4. Thirty-six of the cells were also studied during unexpected descent of the ipsilateral forelimb and six (17%) displayed an increase in discharge (onset latency ca 35 to ca 80 ms); three of these were among those that also responded to contralateral descents. 5. These findings for skilled locomotion requiring a high degree of visuomotor coordination are discussed and it is concluded that the motor cortex is rapidly informed regarding unexpected perturbations delivered to the contralateral forelimb at the onset of stance and that changes are evoked in the pattern of impulse traffic descending via the pyramidal tract.