Visual Control of Postural Orientation and Equilibrium in Congenital Nystagmus

Abstract

purpose. To investigate how humans with congenital nystagmus (CN) use visual information to stabilize and orient their bodies in space.

methods. Center of foot pressure (COP) and head displacements in the lateral plane were recorded using a sway platform and Schottky barrier photodetector, respectively. In experiment 1, a comparison was made of the oscillatory characteristics of body sway with eyes open compared with eyes closed. Experiment 2 studied the postural readjustments made in response to absolute or relative motion (motion parallax) of objects in the visual scene, generated by lateral displacement of background scenery.

results. Experiment 1 revealed that subjects with CN were not able to use visual information to stabilize COP but were able to stabilize the head at frequencies lower than 1 Hz. Experiment 2 showed that in response to the displacement of a visual display, for both absolute motion and motion parallax, subjects with CN reoriented their body in space in a manner similar to control subjects.

conclusions. The results suggest that despite involuntary eye movements, subjects with CN use orientation cues to control their posture, but not dynamic cues useful to control the rapid oscillations that are particularly important at the level of COP. These findings suggest that in CN, visual control of posture is restricted by low-frequency sampling of the visual scene.

Congenital nystagmus (CN) is an ocular motor disorder characterized by involuntary oscillatory eye movements that disrupt foveal target fixation. The onset of nystagmus occurs during the first 6 months of life. ¹ Typically, the nystagmus is in the horizontal plane, and each cycle is initiated by slow phases with exponentially increasing velocity, taking the eyes off the target. A fast phase returns the eyes to the object of regard, at which time there is usually a period of transient stability: a foveation period of 10 to more than 100 msec duration within which the target is visually sampled. ² Despite almost constant eye movement, oscillopsia, or impairment of visuomotor coordination is rare. 

Visual information is an essential factor in the multisensory control of movement and balance. ³ Unlike vestibular information, visual motion signals are relativistic: that is, a displacement of the subject or an external object can yield similar patterns of retinal motion stimuli. Therefore, visual control of posture depends on the ability to differentiate optical flow due to self-motion from that due to object motion. Such a differentiation could be compromised in individuals with involuntary eye movements such as CN. Accordingly, in addition to increased thresholds for motion detection, ^{4 5 6 7} individuals with CN make less use of visual information to control stance, measured as center of foot pressure (COP), than do control subjects. ⁷  

We present detailed investigations of the dynamics of visual postural control in subjects with CN. The first experiment sought to identify the frequency bands in which visual control of equilibrium (the oscillatory component of body sway) is disturbed by CN. Because oscillations of the COP have a higher frequency range than upper parts of the body ^{8 9} both COP and head oscillations were recorded. The second experiment determined the extent to which subjects with CN reorient (i.e., tilt or displace in a given direction) their bodies in space in response to visual flow. Normal subjects lean in the direction of motion ^{10 11 12} when viewing a moving background. This postural adjustment reverses in direction when a stationary target providing motion parallax cues is placed between the standing subject and the moving background. ^{13 14} Because of the intrinsic visual instability and raised thresholds for motion detection in CN, we hypothesized that these subjects may have disturbed patterns of postural response to such visual motion stimuli. 

Materials and Methods

Subjects

Nine adult subjects with horizontal CN ¹⁵ and Snellen visual acuity of at least 0.33 in one eye were studied. Six took part in experiment 1 (subjects 1–6 in Table 1 ), and all participated in experiment 2 (subjects 1–9 in Table 1 ). Fourteen healthy, age-matched subjects participated in experiment 1. Six of these control subjects (mean age, 38.5 years) had their visual acuities artificially reduced to match those of the subjects with CN (acuity control group). These subjects wore spectacles with fogging lenses (×1 magnification lenses with ground surfaces) for 15 minutes before the beginning of the experiment. Because two subjects with CN had extremely low vision in one eye (<0.1), two control subjects were tested monocularly in addition to fogging. The other eight subjects (mean age, 40.5 year) were tested with normal vision (normal control group). Fourteen healthy adults, age matched to the subjects with CN (mean age, 38 years) and having normal Snellen acuity, took part in experiment 2. Written informed consent was obtained according to the Declaration of Helsinki. 

Postural Recordings

The studies focus on postural adjustments in the lateral plane in response to lateral movements of both the visual scene and the eyes, because a number of published studies as well as our pilot studies have shown that postural responses are coplanar with the visual motion stimuli deployed. ^{12 16 17 18} Postural movements in other directions are uncorrelated. In both experiments, subjects stood barefoot on a rigid board placed on top of a slab of foam rubber (height, 5 cm; specific weight, 30 g/dm³) resting on a sway platform that transduced postural sway as displacement of COP in the lateral direction. Feet were splayed at 30° with heels together. The foam increased the instability of the subjects so that the effect of vision on sway was enhanced. ^{19 20} Subjects also wore a lightweight helmet, on top of which was mounted an infrared light-emitting diode. Displacements of the light were transduced by imaging a top view of the subject onto a two-dimensional Schottky barrier photodetector (United Detector Technologies, Hawthorne, CA) situated 40 cm above the head. Light from the diode was projected onto the detector surface through a Mamiya medium format lens (45-mm focal length, f 2.8; Mamiya America Corporation, Elmsford, NY) mounted on the front of the photodetector. Use of this technique to record complex human movements has been described previously. ²¹ One axis of the detector was oriented in the lateral plane to transduce lateral head sway. The resolution of the detector in the configuration deployed was of 0.1 mm, linear up to ±8 cm. Head sway values were normalized for each subjects’ height (signal × mean height of group/individual’s height). All signals were filtered with a band-pass filter of less than 30 Hz and thereafter digitally sampled at 125 Hz. 

Eye Movement Recording

Bitemporal, direct coupled, electro-oculographic recordings of horizontal eye movements were performed on all subjects with CN during experiments 1 and 2, to monitor the predominant fast phase direction and the frequency of the nystagmus. 

Experiment 1: Visual Control of Equilibrium

Subjects were tested under two conditions: eyes open and eyes closed. In both conditions, subjects were instructed to stand still and relax with hands at the sides. With eyes open, subjects were asked to fixate a small cross (1 × 1 cm) placed at the center of an earth-fixed window frame (30 × 24 cm) at a distance of 50 cm from the eyes. The room was normally illuminated. The order of presentation of the eyes-open, eyes-closed conditions was counterbalanced. Each condition was presented for 1 minute, from which the last 50 seconds were analyzed. 

Postural equilibrium in the lateral direction was evaluated as sway path and also in the frequency domain. The sway path is the length of the path described by the COP or the head and is defined as the sum of the distances between sequential points sampled during the analysis period (50 seconds). To calculate power spectra, the 50-second epochs were detrended using a line of best fit and then windowed with a Hanning function. A fast Fourier transform algorithm was then applied to the entire 6250-point signal (Matlab; The Mathworks, Natick, MA). The resultant power spectrum had a frequency resolution of 0.02 Hz and bandwidth from 0 to 62.5 Hz. For subsequent analysis, only the 0- to 5-Hz range was considered. For statistical analysis the frequency components from 0.02 to 5 Hz were grouped into 20 bands, each spanning 0.25 Hz with 12 or 13 (alternating) discrete frequency components per band. The power in each of these bands was then calculated by summing the frequency components. 

Analysis of variance (ANOVA) was used to investigate the effects of vision in the three subject groups. A two-factor design was used for sway path analysis (3 × 2) with group as the between-subject factor (CN, acuity controls, and normal controls) and vision as the within-subject factor (eyes closed versus eyes open). The Tukey test was used for post hoc comparison. A three-factor design was used for spectral analysis (3 × 2 × 20), with frequency (0–0.25 Hz to 4.75–5 Hz) as a second within-subject factor. 

Experiment 2: Visually Induced Body Sway

Postural reorientation was provoked by moving background visual scenery. The scenery was a flat board (2 × 3 m) subtending 67° height × 90° width of visual angle, oriented in the transverse plane, 150 cm from the subject. Photoluminescent yellow-green stripes fixed to the board defined the outline of a house (Fig. 3) in an otherwise dark room. The board was mounted on a motorized wheeled chassis running on a linear track. For background motion, the board first accelerated for 1.25 seconds, rightward or leftward, and then maintained a constant velocity of 6 cm/sec for 8.5 seconds. The overall displacement was 58 cm, subtending 21° from the subject’s viewing position. 

Test conditions: In condition 1, subjects were asked to keep looking straight ahead at the background, which was the only object in the visual scene, and not to follow any particular point (absolute motion, Fig. 3A ). In condition 2, subjects fixated a cross (1 × 1 cm) in the center of a foreground target consisting of a photoluminescent rectangular window frame lattice (30 cm wide, 24 cm high). This window was located straight ahead of the subject, 50 cm from the eyes and 100 cm in front of the visual background. The background was visible through the window panes (motion parallax, Fig. 3B ). Under each condition, subjects underwent 15 pseudorandomized trials: 5 with background motion to the right, 5 with motion to the left, and 5 control trials with the background stationary. 

Postural reorientation in the lateral direction was evaluated as the shift in the average position of the COP and of the head during the constant-velocity part of stimulus motion, relative to a 4-second baseline preceding background motion. Trials were averaged for each subject and visual condition. Student’s t-tests were used to compare CN with control subject data. 

Results

Experiment 1

The sway path data measured from COP and head recordings for CN and the two control groups are given in Table 2 . 

Effect of Vision on COP

Comparisons between the sway path lengths measured in the eyes-open and eyes-closed conditions indicated that visual stabilization of the COP was more effective in the two control groups than in the subjects with CN. The sway path of the COP with eyes open was reduced only by 19% in the subjects with CN compared with 59% and 57% in the acuity control and normal control subjects, respectively. Figure 1A illustrates the improvement of stability (%) with vision for each individual CN and acuity control subject. The significant interaction between the group factor (CN, acuity control and normal control groups) and the vision factor (eyes closed versus eyes open) confirmed that vision was more stabilizing in the two controls groups compared with the CN group (ANOVA: P < 0.05). Post hoc comparison indicated that there was no difference among the three groups with eyes closed (P > 0.05). In the eyes-open condition, the sway path was longer in subjects with CN than in the other two control groups (P < 0.01). No differences of any kind were observed between the acuity control and normal control subjects (P > 0.05). 

Spectral analysis of COP for the CN and the acuity control subjects is shown in Figure 2 . The frequency characteristics of the visual effect on COP displacements can be inferred from a comparison of the spectra of sway obtained with eyes closed versus eyes open. ¹² In CN, visual stabilization of the COP was restricted to frequencies lower than 1 Hz (Fig. 2A) . In contrast, for acuity control subjects (and normal control subjects), vision had an effective stabilizing influence on COP throughout the frequency range 0 to 5 Hz (Fig. 2B for acuity control subjects). Visual stabilization of the COP was significantly more effective in the control subjects (acuity control and normal control subjects) than in the subjects with CN, reflected by the significant interaction between the group factor and the vision factor (ANOVA: P < 0.05). ANOVAs examining for within group effects indicated that the effect of vision in the subjects with CN failed to reach significance either as a main effect (P = 0.11) or in interaction with the frequency factor (P = 0.58). The effect of vision was similar in the two control groups with a significant main effect of vision (P < 0.05) and no interaction with the frequency factor (P > 0.05). 

Effect of Vision on Head Sway

Comparisons between the eyes-open and the eyes-closed conditions indicated that the sway path length was shorter in the three groups of subjects when their eyes were open than when eyes were closed (see Table 2 ). The sway path length with vision was reduced by 27% in the subjects with CN, 34% in the acuity control subjects, and 38% in the normal control subjects, compared with eyes closed. As can be seen in Figure 1B , the improvement in stability with vision was similar to that of acuity control subjects in five of the six subjects with CN. Statistical analysis (ANOVA) confirmed that the reduction of the sway path length with vision was similar in the three groups of subjects. Post hoc comparisons indicated that there was no difference among the three groups in the eyes-closed condition (P > 0.05). With eyes open a significant difference between the CN and the normal control subjects was observed (P < 0.05), but no other comparison reached significance. 

Spectral analysis showed that visual stabilization of the head in the three groups primarily affected low frequencies of head movement. In the subjects with CN (Fig. 2C) visual stabilization of the head was restricted to less than 1 Hz, whereas in the two control groups, visual stabilization of the head was apparent up to 2 to 3 Hz (Fig. 2D for acuity-control subjects). Although visual stabilization of the head appeared to have a higher frequency dynamic in the acuity control and normal control groups, the magnitude of visual stabilization of head sway was similar in the three groups. This was shown by the absence of significant interactions in ANOVAs examining interactions among group, vision, and frequency factors (P > 0.05). The stabilizing effect of vision was significant in the three groups of subjects, as a main effect or in interaction with the factor frequency (P < 0.05). 

Experiment 2

Postural responses to leftward and rightward stimuli were always of similar amplitude, and thus the data were combined (Table 3) . Figure 3 shows sample records of head displacements for a subject with CN during background motion, with both absolute motion and motion parallax. With absolute motion, the displacement of the background induced a displacement of the head (and COP) in the same direction as the background, followed by a return to baseline posture on cessation of the stimulus. In both the CN-affected and control subjects, a postural adjustment in the direction of motion was observed in response to absolute motion, with a similar amplitude in the two groups, both for the COP (t = 0.13, P > 0.05) and for the head (t = 0.21, P > 0.05). 

With a foreground target (i.e., motion parallax) a shift of head position (and COP) in the direction opposite to stimulus motion was induced (Fig. 3) . These postural adjustments in the direction opposite to background motion were significant departures from baseline and were of similar amplitude in the two subject groups (COP: t = 0.27, P > 0.05; head: t = 0.38, P > 0.05). 

An additional analysis was made of the results from the seven subjects with CN who had sustained, unidirectionally beating nystagmus (Table 1) to test whether the direction of nystagmus affected the postural readjustment to leftward and rightward stimuli. The postural responses were inverted in the two subjects with CN with right-beating nystagmus to make their data comparable with that of the other five subjects with CN with left-beating nystagmus. Student’s t-tests comparing response amplitudes in the same versus opposite direction to the nystagmus fast phase showed that the nystagmus direction had no effect on COP and head data, either for absolute motion or motion parallax (P > 0.05). 

Discussion

Experiment 1 showed that visual control of equilibrium in the lateral direction in subjects with CN appears to have greatest efficacy for low-frequency components of sway (<1 Hz). However, visual control of equilibrium was less effective in subjects with CN than in control subjects. We could detect a marginal loss of visual stabilization of high-frequency head movements in subjects with CN compared with control subjects. However, high-frequency components to head movement were minimal in both the CN-affected and the control subjects (acuity and normal controls) and thus had little implication for postural stability as measured. COP had more power at high frequencies than head movement. The reduction of COP instability due to vision, was smaller in subjects with CN than in control subjects across all frequencies, but particularly at frequencies of more than 1 Hz. No difference was observed between the two control groups tested (acuity controls and normal controls), indicating that the differences between the subjects with CN and the control subjects was not a consequence of the slightly reduced visual acuity of the subjects with CN included in experiment 1. With eyes closed, COP and head stability were similar in the subjects with CN and the control subjects. These results are consistent with previous observations, ⁷ which support the thesis that somatosensory and vestibular controls of posture in subjects with CN are normal. 

Experiment 2 showed that the use of visual information to control body orientation in space (i.e., overall tilt) was normal in CN. Visually induced body sway under conditions of absolute motion and motion parallax did not differ among subject groups. Consistent with reports in the literature, absolute motion induced ipsidirectional body sway, ^{10 11 12} whereas juxtaposing a stationary target between subject and background provoked sway contradirectional to background motion. ^{13 14} Thus, despite their nystagmus, subjects with CN made normal use of visual motion cues, including motion parallax, to control postural orientation. 

Insight into the impairment of visual control of high-frequency postural instability in CN is given by the behavior of normal subjects in stroboscopic light, when the flashes are presented at a strobing frequency of 3 to 5 Hz. ^{22 23 24 25 26 27} This frequency range is similar to the nystagmus frequency in this sample of subjects with CN (see Table 1 ). Isableu et al. ²² showed that subjects with normal vision, standing in front of a tilted frame, leaned in the direction of tilt under normal or stroboscopic lighting (2.8 Hz). Displacement of an oscillating background under strobed light also causes a continuous modulation of low-frequency postural sway. ²³ These results indicate that discrete visual sampling is sufficient for controlling body orientation. However, unlike body orientation, normal equilibrium appears to be degraded under stroboscopic vision at frequencies lower than 6 Hz. ^{24 25 26} Measured at different levels, from ankle to head, the destabilizing effect of such discrete visual sampling principally affects the lower parts of the body. ²⁷ The latter investigators concluded that discrete visual information (static cues) were sufficient to control the upper part of the body, which has predominantly low-frequency dynamics. Visual motion cues (dynamic cues) control oscillations of the lower part of the body, which extend through a higher frequency range. Thus, subjects with CN appear to be similar to normal subjects in stroboscopic light, in that they share some ability to orient and control low-frequency head instability but are less able to control the higher frequency instabilities of the COP with the visual cues available. Dynamic visual cues, requiring continuous visual feedback, appear to be particularly crucial for fast stabilization of the COP. 

The similarity between normal subjects in stroboscopic light and subjects with CN is consistent with the concept that the waveform of CN affords intermittent, low-frequency visual sampling at the time of the foveation periods. Subjects with CN do not behave as though they were exposed to continuous visual motion because of their nystagmus, and this intermittent sampling of vision to control posture may be related to the mechanism whereby they suppress oscillopsia. 

The mechanisms proposed for suppressing oscillopsia in CN include a reduced sensitivity to retinal image motion ^{5 6 28} ; an ability to extract visual information during foveation periods (the parts of the nystagmus waveform when the eyes are quiescent and images are most stable on the fovea) and to ignore the smeared vision during high-velocity slow phases ^{2 29 30} ; and the use of extraretinal signals—i.e., efference copy of the CN waveform—to negate the visual effects of the oscillation. ^{4 31 32} Of these, the most recent evidence suggests that the efference copy of the CN waveform is the major factor in oscillopsia suppression. ^{4 32} Although foveation periods may not be primarily responsible for oscillopsia suppression, they are important for visual acuity, ³² and they may be responsible for the discrete sampling of visual cues to postural orientation in CN. 

 Submitted for publication October 18, 1999; revised April 4 and June 16, 2000; accepted July 5, 2000.

 Commercial relationships policy: N.

 

Table 1.

 

Clinical Details of the Nine Subjects with CN Who Took Part in the Experiment

Subject	Age/Sex	Snellen Acuity	Frequency (Hz)	Stereopsis (seconds of arc)	Dominant Waveform
1	38 M	0.5; 0.5	4.5	50	Jerk left
2	40 F	<0.1; 0.5	4	Nil	Alternating
3	57 M	0.67; 0.67	4	80	Pendular
4	33 M	0.33; <0.1	4	Nil	Jerk left
5	24 F	0.17; 0.67	4.5	Nil	Jerk right
6	41 M	0.67; 0.67	5	140	Jerk left
7	30 M	0.33; 0.33	4	800	Jerk right
8	30 M	0.33 ;0.33	3.5	Nil	Jerk left
9	56 M	0.67; 0.5	4	100	Jerk left

 Experiment 1: subjects 1–6; experiment 2: subjects 1–9.

Table 2.

 

Mean Sway Path Length and SD of Both the COP and the Head for the Eyes–Closed and Eyes–Open Conditions in CN and Acuity and Normal Control Groups

	Eyes Closed			Eyes Open
		M	SD	M	SD
Head
CN	67.8	13.2	49.1	12.1
Acuity control	56.5	12.2	35.9	9.5
Normal control	53.2	17.5	31.9	13.3
COP
CN	115.1	15.6	92.4	23.1
Acuity control	111.3	23.9	45.8	10.1
Normal control	128.1	55.1	50.3	11.5

 Data are in centimeters.

Figure 1.

 

Percentage improvement of stability with vision calculated for each subject with CN (○) and acuity control (♦) subject for both the COP (A) and the head (B) in experiment 1. This index of performance was computed as: [(eyes-closed score − eyes-open score)/(eyes-closed score)] × 100.

Figure 2.

 

Average power spectra of displacement of the COP (A, B) and the head (C, D) in CN and acuity control subjects in experiment 1. Power spectral density is in log₁₀ (in square centimeters with a frequency resolution of 0.02 Hz). Error bars are the SD for each frequency band. Note that the decay in power with increasing frequency for oscillatory head movement (C, D) was greater than for oscillations of the COP.

Table 3.

 

Average Position and SD of the COP and the Head under Conditions of Absolute Motion and Motion Parallax, with Moving or Stationary Background in Subjects with CN and Control Subjects

	Moving Background							Stationary Background
		Absolute Motion			Motion Parallax			Absolute Motion			Motion Parallax
		M	SD	M	SD	M	SD	M	SD
Head
Control	3.90	4.1	−5.48	2.9	0.42	1.4	0.48	1.1
CN	3.54	4.2	−6.15	4.5	−0.26	3.9	0.43	2.1
COP
Control	2.92	2.4	−3.73	1.7	0.3	0.6	0.4	1.01
CN	2.79	2.5	−4.01	3.04	0.4	2.9	−0.25	0.85

 Data are expressed in millimeters.

Figure 3.

 

Sample records of head displacements (top traces) of a subject with CN under conditions of absolute motion and motion parallax during background motion. Upward deflections indicate deviation in the direction of stimulus motion. The drawings show the setups for experiment 2.

The authors thank the volunteers who enthusiastically participated in this project.