Head oscillations in infantile nystagmus syndrome Fatema F. Ghasia, MD,a and Aasef G. Shaikh, MD, PhDb PURPOSE METHODS
RESULTS
CONCLUSIONS
To quantitatively characterize eye and head oscillations in patients with infantile nystagmus syndrome (INS). Vertical and horizontal eye and head position in INS patients were measured simultaneously at a sampling frequency of 500 Hz. Eye and head movements were measured continuously for 180 seconds. The data was calibrated and converted to angular vectors, which were further analyzed with custom software. A total of 10 patients with INS were included: 3 with pseudo-jerk, 3 with pure-jerk, 2 with pseudo-pendular with foveating saccade form of jerk, 1 with bidirectional jerk, and 1 with asymmetric pendular nystagmus waveforms. None of the patients had periodic, aperiodic, or a superimposed latent nystagmus component. Two types of head oscillations were observed: one with a frequency of 1–3 Hz, present in all patients; and another with a frequency range of 5–8 Hz, present in only 7 patients. High-frequency oscillations were episodic, whereas low-frequency oscillations were constantly present. Peak velocity of the high-frequency head oscillations and eye velocity of nystagmus were not correlated, suggesting that these oscillations did not influence foveation. Two types of head oscillations were found in INS patients: a constant, low-frequency and an episodic, high–frequency. Lack of correlation between the foveation period of nystagmus and peak head velocity during high-frequency oscillations suggests a coexisting pathological phenomenon rather than a compensatory mechanism used to improve the visual acuity. ( J AAPOS 2015;19:38-41)
I
nfantile nystagmus syndrome (INS), spasmus nutans, and fusion maldevelopment nystagmus are characteristically first observed in childhood.1 Head oscillations often accompany nystagmus in these syndromes.2-4 Lowfrequency head oscillations at 2–3 Hz frequency have been reported in INS.4-6 The role of such low-frequency head oscillations in INS is controversial. Some have suggested that they may represent an adaptive strategy for minimizing nystagmus,5-7 whereas others have suggested that they are due to pathologic reverberations in the cephalo-motor neural circuit.4 We describe two types of head oscillations observed in 10 INS patients. Consistent with the previous literature, the first type had low oscillation frequency in the 1–3 Hz range.4-6 The second, novel subtype featured high-frequency oscillations in the 5–8 Hz range.
Materials and Methods We invited INS patients visiting the Cole Eye Institute between September 2012 and March 2014 to participate in the study. Author affiliations: aCole Eye Institute, Cleveland Clinic, Cleveland, Ohio; bDepartment of Neurology, Cleveland Clinic, Cleveland, Ohio Submitted August 7, 2014. Revision accepted October 13, 2014. Correspondence: Fatema F. Ghasia, MD, Cole Eye Institute, Cleveland Clinic Foundation, 2022 E 105th Street, Cleveland, OH 44106 (email: fatemaghasia@gmail. com). Copyright Ó 2015 by the American Association for Pediatric Ophthalmology and Strabismus. 1091-8531/$36.00 http://dx.doi.org/10.1016/j.jaapos.2014.10.024
38
The experiment protocols were approved by the Institutional Review Board of the Cleveland Clinic and conformed to guidelines of the US Health Insurance Privacy and Accountability Act of 1996. All subjects provided written informed consent. The eye and head movements of INS patients were measured and analyzed as described below. Eye oscillations were first examined clinically in primary gaze, eccentric gaze position at near and distance, under monocular and binocular conditions. During prolonged clinical examination none of the patients experienced an aperiodic, periodic, or superimposed latent nystagmus component. Ocular vergence performed during clinical examination dampened the nystagmus. Neurological examination was otherwise normal.
Experiment Protocol A high-resolution video-based eye tracker (EyeLink 1000; SR Research, Ontario, Canada) was used to measure horizontal and vertical eye positions under monocular and binocular viewing conditions. The recordings were obtained for 180 seconds to determine whether there was any latent component present. The trial duration was sufficient for characterizing types of head oscillations. In addition, horizontal and vertical head positions were measured under conditions of binocular viewing. Both horizontal and vertical eye and head positions were measured at temporal resolution of 500 Hz and a spatial resolution of 0.01 . Subjects were seated at 55 cm from the center of the LCD screen on which the visual target was displayed. A red visual target was projected straight ahead and at 5 and 10 in the rightward, leftward, upward, and downward directions. Subjects were instructed to fix their gaze on the target projected on
Journal of AAPOS
Volume 19 Number 1 / February 2015
Ghasia and Shaikh
39
Table 1. Information on range of refractive errors found within the study group Type of refractive error
Number
Mean SE, D
Standard deviation, D
Range of SE, D
Mean initial visual acuity, logMAR
Hypermetropia Astigmatism Hypermetropia and astigmatism Myopia
13 12 8 0
5.19 1.43 6.20 -
1.06 1.11 1.45 -
4.00-7.50 0.00-3.00 3.50-8.13 -
0.553 0.452 0.591 -
D, diopter; SE, spherical equivalent. the LCD screen with a white background in a completely dark room. A laser was mounted on a headband. The subjects were asked to align the head-mounted laser beam with the visual target on a computer screen. The eye and head position data for known size of visually guided isolated eye and isolated head movements were used to calibrate head and eye position signal. Three trials were performed. Further analysis was performed offline on calibrated angular position vectors.
Data Analysis Epochs of head position signal with low- and high-frequency oscillations were interactively selected and separately analyzed. Both authors independently performed the analysis; the same author performed this analysis twice on separate occasions. Examiners were looking for epochs of high-frequency oscillations that were easily identifiable from the background low-frequency oscillations. Intra- and interexaminer consistency was excellent. Frequency and amplitude of head oscillations were computed for each group. A cycle-by-cycle analysis was performed on two-dimensional composite vectors (e-Supplement 1, available at jaapos.org). The x-coordinates of the intersection of the trace with the abscissa (moving from a negative to a positive value) were recorded. The xcoordinate of the first data point that crossed the abscissa marked the beginning of a cycle, and the subsequent data point marked the end. The cycle width was computed by measuring the difference between two crossings. The inverse of the cycle width yielded the cycle frequency and the difference between the peak and trough yielded the cycle amplitude. Data was analyzed with custom software in MATLAB (Mathworks, Natick, MA). The wavelet toolbox was used to assess the effects of high-frequency head oscillations on the power of irregular eye oscillations. The wavelet transformation is comparable to the Fourier transformation; however, wavelets localize in both time and frequency domains, whereas the Fourier transform localizes to the frequency domain alone. The wavelet power determines the distribution of the energy within the data array; the wavelet power is equivalent to the amplitude squared.
Results We measured eye and head movements in 10 INS patients (5 females), with a mean age of 30.2 19.6 (range, 11-66 years). Demographic and clinical characteristics are summarized in Table 1. We found two types of head oscillations. One type was a low-frequency (1–3 Hz), constant oscillation; the other type, a high-frequency episodic oscillation (5–8 Hz). Figure 1 depicts an example of both types in a single INS patient. Figure 1A depicts horizontal and
Journal of AAPOS
FIG 1. Example of eye and head oscillations in an INS patient. The first two traces depict horizontal and vertical eye positions. The positions are plotted on y-axis; time, on the x-axis. Corresponding horizontal and vertical head oscillations are plotted in third and fourth traces, respectively. The arrows indicate examples of head oscillations; the first arrow from the left show low frequency oscillations, while the second arrow from the left depicts high-frequency oscillations.
vertical eye positions; Figure 2B, horizontal and vertical head positions. The eye position trace shows a primarily horizontal jerk nystagmus with a superimposed small vertical component. These oscillations have distinct trajectory, which is typical of INS waveforms.8,9 The head position plot shows two discrete types of oscillations. The first is the low-frequency type that is primarily in the horizontal plane with a small component in the vertical axis. It is present during the entire recording. The second is the high-frequency type, prominent in the vertical axis. The high-frequency head oscillations occurred as burst of brief episodes superimposed on the low-frequency head oscillations. The low-frequency head oscillations were seen in all 10 subjects, whereas the high-frequency head oscillations were seen in 7 of 10 subjects. We quantified the frequency and amplitude of these head oscillations in all subjects collectively. The box-andwhisker plot of Figure 2A depicts the summary of the frequency of fast and slow head oscillations. The median frequency and standard deviation of fast head oscillations
40
Ghasia and Shaikh
Volume 19 Number 1 / February 2015
FIG 3. Traces of histograms summarizing the distribution of the angle of the trajectory of high- and low-frequency head oscillations. The black trace shows the distribution of the direction of the trajectory of highfrequency oscillations; the gray line, low-frequency oscillations. FIG 2. Box-and-whisker plots depicting the summary of cycle-bycycle frequency (A) and amplitude (B) of high- and low-frequency head oscillations from all 7 patients. The notch depicts the 95% confidence interval; the horizontal line in the center of the notch, the median value; the length of the box, the interquartile length; and plus symbols, outliers.
was 5.84 0.2 Hz; of slow head oscillations, 1.83 0.12 Hz. The difference in median frequencies between the two types of oscillations was statistically significantly different (t test; P \ 0.0001). The box-and-whisker plot of Figure 2B depicts the summary of amplitude of fast and slow head oscillations. The median amplitude of the high-frequency head oscillations was 0.74 0.11 whereas the median amplitude of the low-frequency head oscillations was 1.57 0.35 . The difference in median amplitude between the two head oscillations was statistically significantly different (t test; P \ 0.001). Figure 3 shows the frequency and direction of slow and fast head oscillations. The mean angle of trajectories of high-frequency head oscillations was 63.2 23.9 ; of the low-frequency head oscillations, 33.3 33.26 . The distribution of trajectories of the two types of head oscillations differed significantly (two-tailed K-S test; P 5 0.03). If the high-frequency oscillations were adaptive, and thus actively damping the nystagmus, then they would change the kinematic properties (ie, frequency contents and amplitude) of the nystagmus waveform. Therefore, we compared the nystagmus waveform during the epochs of high-frequency head oscillations to the nystagmus waveform when high-frequency head oscillations were absent. Because high-frequency oscillations were present in multiple short epochs, we collated data from all epochs in a given patient and used wavelet analysis to assess the frequency content of the complex waveform. Continuous
wavelet transform is helpful while analyzing irregular signals such as that of INS waveform because it offers time to frequency localization. Figure 4A depicts a continuous wavelet transformation of nystagmus waveform depicted in Figure 4B. In Figure 4A the y-axis depicts the wavelet level (a unit comparable to the frequency), the x-axis represents the time, and the color scale (color on the Web) depicts wavelet power (a unit comparable to the intensity). In most instances, the maximum wavelet power in a given epoch of oscillations remains between 105 and 144. We sampled the power at all wavelet levels and constructed the cumulative sum of the wavelet power spectrum. We computed and compared wavelet power spectra for eye nystagmus when slow-frequency and high-frequency oscillations were present. Figure 4C depicts the summary of normalized cumulative sum of wavelet power spectrum from all patients. Both the solid and the dashed lines of identical color superimpose in each subject. Hence, there was a resemblance in the frequency content and the power distribution in the given frequency of the waveform. This subjective resemblance was statistically proven with Kolmogorov-Smirnov statistics. Hence we concluded that rapid head oscillations did not affect the nystagmus and that high-frequency head oscillations were not adaptive but a coexisting pathological process.
Discussion To our knowledge, ours is the largest case series of INS patients with head oscillations in which head and eye movements have been measured and quantified. Our results reveal two types of head oscillations in these patients: lowfrequency head oscillations, similar to ones previously described in smaller case series,4,5 and high-frequency head oscillations. Most previous reports of low-frequency
Journal of AAPOS
Volume 19 Number 1 / February 2015
Ghasia and Shaikh
41
FIG 4. Continuous wavelet transform (A) of INS waveform (B). In the comparison of normalized cumulative power spectrum of eye nystagmus in the presence of high-frequency oscillations (C), each of 7 patients is depicted by a shade of gray: the dashed line represents power spectrum in the presence of high-frequency oscillations; solid lines depict power spectrum in the absence of the high-frequency oscillations.
oscillations measured eye and head movements simultaneously and found that head oscillations do not influence the eye movements in INS.4,5,10 These studies proposed that the low-frequency head oscillations in INS represent a pathologic tremor rather than an adaptive strategy to damp the nystagmus.4 More recently, Anagnostou and colleagues6 suggested that the vertical head movements in INS patient could be a compensatory phenomenon and showed that although horizontal head movements had no effect on the nystagmus waveforms, the vertical head oscillations changed the nystagmus waveform and were associated with decreased frequency and transient decrease in slow-phase velocity, thus allowing for a longer foveation period. Furthermore, the frequency of the vertical head oscillation found by Anagnostou and colleagues6 was similar to that of the eye oscillations. Anagnostou and colleagues6 offered their results with an important caveat, however; the diagnosis of INS was questionable, and there was a possibility of spasmus nutans. In contrast, our study found that high- and low-frequency head oscillations were present in INS patients with or without albinism. Moreover, our patients had a definitive diagnosis of INS. Jan and colleagues7 also described head oscillations in visually impaired children, 4 with albinism and 2 with congenital motor nystagmus. Their study included subjects who appeared clinically to have head oscillations on attempted visual fixation, but the oscillation frequency was 1–2 Hz, as reported in previous studies.4,5,10 Our results show that low-frequency head oscillations are primarily in the horizontal direction and that episodic bursts of high-frequency head oscillations are predominantly vertical. These high-frequency vertical head oscillations did not appear to affect the nystagmus waveforms.
Journal of AAPOS
This suggests that the vertical high-frequency head oscillations superimposed on the slow-frequency horizontal head oscillations in INS patients might represent pathological reverberations in the cephalomotor system leading to the high-frequency head tremor. References 1. Hertle R, CEMAS Working Group. A National Eye Institute Sponsored Workshop and Publication on the Classification of Eye Movement Abnormalities and Strabismus (CEMAS). The National Eye Institute Publications, http://www.nei.nih.gov; 2001. 2. Brodsky MC, Wright KW. Infantile esotropia with nystagmus: a treatable cause of oscillatory head movements in children. Arch Ophthalmol 2007;125:1079-81. 3. Gresty M, Leech J, Sanders M, Eggars H. A study of head and eye movement in spasmus nutans. Br J Ophthalmol 1976;60:652-4. 4. Carl JR, Optican LM, Chu FC, Zee DS. Head shaking and vestibuloocular reflex in congenital nystagmus. Invest Ophthalmol Vis Sci 1985;26:1043-50. 5. Gresty MA, Halmagyi GM. Abnormal head movements. J Neurol Neurosurg Psychiatry 1979;42:705-14. 6. Anagnostou E, Spengos K, Anastasopoulos D. Single-plane compensatory phase shift of head and eye oscillations in infantile nystagmus syndrome. J Neurol Sci 2011;308:182-5. 7. Jan JE, Groenveld M, Connolly MB. Head shaking by visually impaired children: a voluntary neurovisual adaptation which can be confused with spasmus nutans. Dev Med Child Neurol 1990;32: 1061-6. 8. Dell’Osso LF, Daroff RB. Congenital nystagmus waveforms and foveation strategy. Doc Ophthalmol 1975;39:155-82. 9. Abadi RV, Dickinson CM. Waveform characteristics in congenital nystagmus. Doc Ophthalmol 1986;64:153-67. 10. Gresty M, Halmagyi GM, Leech J. The relationship between head and eye movement in congenital nystagmus with head shaking: objective recordings of a single case. Br J Ophthalmol 1978;62: 533-5.
RESULTS
CONCLUSIONS
To quantitatively characterize eye and head oscillations in patients with infantile nystagmus syndrome (INS). Vertical and horizontal eye and head position in INS patients were measured simultaneously at a sampling frequency of 500 Hz. Eye and head movements were measured continuously for 180 seconds. The data was calibrated and converted to angular vectors, which were further analyzed with custom software. A total of 10 patients with INS were included: 3 with pseudo-jerk, 3 with pure-jerk, 2 with pseudo-pendular with foveating saccade form of jerk, 1 with bidirectional jerk, and 1 with asymmetric pendular nystagmus waveforms. None of the patients had periodic, aperiodic, or a superimposed latent nystagmus component. Two types of head oscillations were observed: one with a frequency of 1–3 Hz, present in all patients; and another with a frequency range of 5–8 Hz, present in only 7 patients. High-frequency oscillations were episodic, whereas low-frequency oscillations were constantly present. Peak velocity of the high-frequency head oscillations and eye velocity of nystagmus were not correlated, suggesting that these oscillations did not influence foveation. Two types of head oscillations were found in INS patients: a constant, low-frequency and an episodic, high–frequency. Lack of correlation between the foveation period of nystagmus and peak head velocity during high-frequency oscillations suggests a coexisting pathological phenomenon rather than a compensatory mechanism used to improve the visual acuity. ( J AAPOS 2015;19:38-41)
I
nfantile nystagmus syndrome (INS), spasmus nutans, and fusion maldevelopment nystagmus are characteristically first observed in childhood.1 Head oscillations often accompany nystagmus in these syndromes.2-4 Lowfrequency head oscillations at 2–3 Hz frequency have been reported in INS.4-6 The role of such low-frequency head oscillations in INS is controversial. Some have suggested that they may represent an adaptive strategy for minimizing nystagmus,5-7 whereas others have suggested that they are due to pathologic reverberations in the cephalo-motor neural circuit.4 We describe two types of head oscillations observed in 10 INS patients. Consistent with the previous literature, the first type had low oscillation frequency in the 1–3 Hz range.4-6 The second, novel subtype featured high-frequency oscillations in the 5–8 Hz range.
Materials and Methods We invited INS patients visiting the Cole Eye Institute between September 2012 and March 2014 to participate in the study. Author affiliations: aCole Eye Institute, Cleveland Clinic, Cleveland, Ohio; bDepartment of Neurology, Cleveland Clinic, Cleveland, Ohio Submitted August 7, 2014. Revision accepted October 13, 2014. Correspondence: Fatema F. Ghasia, MD, Cole Eye Institute, Cleveland Clinic Foundation, 2022 E 105th Street, Cleveland, OH 44106 (email: fatemaghasia@gmail. com). Copyright Ó 2015 by the American Association for Pediatric Ophthalmology and Strabismus. 1091-8531/$36.00 http://dx.doi.org/10.1016/j.jaapos.2014.10.024
38
The experiment protocols were approved by the Institutional Review Board of the Cleveland Clinic and conformed to guidelines of the US Health Insurance Privacy and Accountability Act of 1996. All subjects provided written informed consent. The eye and head movements of INS patients were measured and analyzed as described below. Eye oscillations were first examined clinically in primary gaze, eccentric gaze position at near and distance, under monocular and binocular conditions. During prolonged clinical examination none of the patients experienced an aperiodic, periodic, or superimposed latent nystagmus component. Ocular vergence performed during clinical examination dampened the nystagmus. Neurological examination was otherwise normal.
Experiment Protocol A high-resolution video-based eye tracker (EyeLink 1000; SR Research, Ontario, Canada) was used to measure horizontal and vertical eye positions under monocular and binocular viewing conditions. The recordings were obtained for 180 seconds to determine whether there was any latent component present. The trial duration was sufficient for characterizing types of head oscillations. In addition, horizontal and vertical head positions were measured under conditions of binocular viewing. Both horizontal and vertical eye and head positions were measured at temporal resolution of 500 Hz and a spatial resolution of 0.01 . Subjects were seated at 55 cm from the center of the LCD screen on which the visual target was displayed. A red visual target was projected straight ahead and at 5 and 10 in the rightward, leftward, upward, and downward directions. Subjects were instructed to fix their gaze on the target projected on
Journal of AAPOS
Volume 19 Number 1 / February 2015
Ghasia and Shaikh
39
Table 1. Information on range of refractive errors found within the study group Type of refractive error
Number
Mean SE, D
Standard deviation, D
Range of SE, D
Mean initial visual acuity, logMAR
Hypermetropia Astigmatism Hypermetropia and astigmatism Myopia
13 12 8 0
5.19 1.43 6.20 -
1.06 1.11 1.45 -
4.00-7.50 0.00-3.00 3.50-8.13 -
0.553 0.452 0.591 -
D, diopter; SE, spherical equivalent. the LCD screen with a white background in a completely dark room. A laser was mounted on a headband. The subjects were asked to align the head-mounted laser beam with the visual target on a computer screen. The eye and head position data for known size of visually guided isolated eye and isolated head movements were used to calibrate head and eye position signal. Three trials were performed. Further analysis was performed offline on calibrated angular position vectors.
Data Analysis Epochs of head position signal with low- and high-frequency oscillations were interactively selected and separately analyzed. Both authors independently performed the analysis; the same author performed this analysis twice on separate occasions. Examiners were looking for epochs of high-frequency oscillations that were easily identifiable from the background low-frequency oscillations. Intra- and interexaminer consistency was excellent. Frequency and amplitude of head oscillations were computed for each group. A cycle-by-cycle analysis was performed on two-dimensional composite vectors (e-Supplement 1, available at jaapos.org). The x-coordinates of the intersection of the trace with the abscissa (moving from a negative to a positive value) were recorded. The xcoordinate of the first data point that crossed the abscissa marked the beginning of a cycle, and the subsequent data point marked the end. The cycle width was computed by measuring the difference between two crossings. The inverse of the cycle width yielded the cycle frequency and the difference between the peak and trough yielded the cycle amplitude. Data was analyzed with custom software in MATLAB (Mathworks, Natick, MA). The wavelet toolbox was used to assess the effects of high-frequency head oscillations on the power of irregular eye oscillations. The wavelet transformation is comparable to the Fourier transformation; however, wavelets localize in both time and frequency domains, whereas the Fourier transform localizes to the frequency domain alone. The wavelet power determines the distribution of the energy within the data array; the wavelet power is equivalent to the amplitude squared.
Results We measured eye and head movements in 10 INS patients (5 females), with a mean age of 30.2 19.6 (range, 11-66 years). Demographic and clinical characteristics are summarized in Table 1. We found two types of head oscillations. One type was a low-frequency (1–3 Hz), constant oscillation; the other type, a high-frequency episodic oscillation (5–8 Hz). Figure 1 depicts an example of both types in a single INS patient. Figure 1A depicts horizontal and
Journal of AAPOS
FIG 1. Example of eye and head oscillations in an INS patient. The first two traces depict horizontal and vertical eye positions. The positions are plotted on y-axis; time, on the x-axis. Corresponding horizontal and vertical head oscillations are plotted in third and fourth traces, respectively. The arrows indicate examples of head oscillations; the first arrow from the left show low frequency oscillations, while the second arrow from the left depicts high-frequency oscillations.
vertical eye positions; Figure 2B, horizontal and vertical head positions. The eye position trace shows a primarily horizontal jerk nystagmus with a superimposed small vertical component. These oscillations have distinct trajectory, which is typical of INS waveforms.8,9 The head position plot shows two discrete types of oscillations. The first is the low-frequency type that is primarily in the horizontal plane with a small component in the vertical axis. It is present during the entire recording. The second is the high-frequency type, prominent in the vertical axis. The high-frequency head oscillations occurred as burst of brief episodes superimposed on the low-frequency head oscillations. The low-frequency head oscillations were seen in all 10 subjects, whereas the high-frequency head oscillations were seen in 7 of 10 subjects. We quantified the frequency and amplitude of these head oscillations in all subjects collectively. The box-andwhisker plot of Figure 2A depicts the summary of the frequency of fast and slow head oscillations. The median frequency and standard deviation of fast head oscillations
40
Ghasia and Shaikh
Volume 19 Number 1 / February 2015
FIG 3. Traces of histograms summarizing the distribution of the angle of the trajectory of high- and low-frequency head oscillations. The black trace shows the distribution of the direction of the trajectory of highfrequency oscillations; the gray line, low-frequency oscillations. FIG 2. Box-and-whisker plots depicting the summary of cycle-bycycle frequency (A) and amplitude (B) of high- and low-frequency head oscillations from all 7 patients. The notch depicts the 95% confidence interval; the horizontal line in the center of the notch, the median value; the length of the box, the interquartile length; and plus symbols, outliers.
was 5.84 0.2 Hz; of slow head oscillations, 1.83 0.12 Hz. The difference in median frequencies between the two types of oscillations was statistically significantly different (t test; P \ 0.0001). The box-and-whisker plot of Figure 2B depicts the summary of amplitude of fast and slow head oscillations. The median amplitude of the high-frequency head oscillations was 0.74 0.11 whereas the median amplitude of the low-frequency head oscillations was 1.57 0.35 . The difference in median amplitude between the two head oscillations was statistically significantly different (t test; P \ 0.001). Figure 3 shows the frequency and direction of slow and fast head oscillations. The mean angle of trajectories of high-frequency head oscillations was 63.2 23.9 ; of the low-frequency head oscillations, 33.3 33.26 . The distribution of trajectories of the two types of head oscillations differed significantly (two-tailed K-S test; P 5 0.03). If the high-frequency oscillations were adaptive, and thus actively damping the nystagmus, then they would change the kinematic properties (ie, frequency contents and amplitude) of the nystagmus waveform. Therefore, we compared the nystagmus waveform during the epochs of high-frequency head oscillations to the nystagmus waveform when high-frequency head oscillations were absent. Because high-frequency oscillations were present in multiple short epochs, we collated data from all epochs in a given patient and used wavelet analysis to assess the frequency content of the complex waveform. Continuous
wavelet transform is helpful while analyzing irregular signals such as that of INS waveform because it offers time to frequency localization. Figure 4A depicts a continuous wavelet transformation of nystagmus waveform depicted in Figure 4B. In Figure 4A the y-axis depicts the wavelet level (a unit comparable to the frequency), the x-axis represents the time, and the color scale (color on the Web) depicts wavelet power (a unit comparable to the intensity). In most instances, the maximum wavelet power in a given epoch of oscillations remains between 105 and 144. We sampled the power at all wavelet levels and constructed the cumulative sum of the wavelet power spectrum. We computed and compared wavelet power spectra for eye nystagmus when slow-frequency and high-frequency oscillations were present. Figure 4C depicts the summary of normalized cumulative sum of wavelet power spectrum from all patients. Both the solid and the dashed lines of identical color superimpose in each subject. Hence, there was a resemblance in the frequency content and the power distribution in the given frequency of the waveform. This subjective resemblance was statistically proven with Kolmogorov-Smirnov statistics. Hence we concluded that rapid head oscillations did not affect the nystagmus and that high-frequency head oscillations were not adaptive but a coexisting pathological process.
Discussion To our knowledge, ours is the largest case series of INS patients with head oscillations in which head and eye movements have been measured and quantified. Our results reveal two types of head oscillations in these patients: lowfrequency head oscillations, similar to ones previously described in smaller case series,4,5 and high-frequency head oscillations. Most previous reports of low-frequency
Journal of AAPOS
Volume 19 Number 1 / February 2015
Ghasia and Shaikh
41
FIG 4. Continuous wavelet transform (A) of INS waveform (B). In the comparison of normalized cumulative power spectrum of eye nystagmus in the presence of high-frequency oscillations (C), each of 7 patients is depicted by a shade of gray: the dashed line represents power spectrum in the presence of high-frequency oscillations; solid lines depict power spectrum in the absence of the high-frequency oscillations.
oscillations measured eye and head movements simultaneously and found that head oscillations do not influence the eye movements in INS.4,5,10 These studies proposed that the low-frequency head oscillations in INS represent a pathologic tremor rather than an adaptive strategy to damp the nystagmus.4 More recently, Anagnostou and colleagues6 suggested that the vertical head movements in INS patient could be a compensatory phenomenon and showed that although horizontal head movements had no effect on the nystagmus waveforms, the vertical head oscillations changed the nystagmus waveform and were associated with decreased frequency and transient decrease in slow-phase velocity, thus allowing for a longer foveation period. Furthermore, the frequency of the vertical head oscillation found by Anagnostou and colleagues6 was similar to that of the eye oscillations. Anagnostou and colleagues6 offered their results with an important caveat, however; the diagnosis of INS was questionable, and there was a possibility of spasmus nutans. In contrast, our study found that high- and low-frequency head oscillations were present in INS patients with or without albinism. Moreover, our patients had a definitive diagnosis of INS. Jan and colleagues7 also described head oscillations in visually impaired children, 4 with albinism and 2 with congenital motor nystagmus. Their study included subjects who appeared clinically to have head oscillations on attempted visual fixation, but the oscillation frequency was 1–2 Hz, as reported in previous studies.4,5,10 Our results show that low-frequency head oscillations are primarily in the horizontal direction and that episodic bursts of high-frequency head oscillations are predominantly vertical. These high-frequency vertical head oscillations did not appear to affect the nystagmus waveforms.
Journal of AAPOS
This suggests that the vertical high-frequency head oscillations superimposed on the slow-frequency horizontal head oscillations in INS patients might represent pathological reverberations in the cephalomotor system leading to the high-frequency head tremor. References 1. Hertle R, CEMAS Working Group. A National Eye Institute Sponsored Workshop and Publication on the Classification of Eye Movement Abnormalities and Strabismus (CEMAS). The National Eye Institute Publications, http://www.nei.nih.gov; 2001. 2. Brodsky MC, Wright KW. Infantile esotropia with nystagmus: a treatable cause of oscillatory head movements in children. Arch Ophthalmol 2007;125:1079-81. 3. Gresty M, Leech J, Sanders M, Eggars H. A study of head and eye movement in spasmus nutans. Br J Ophthalmol 1976;60:652-4. 4. Carl JR, Optican LM, Chu FC, Zee DS. Head shaking and vestibuloocular reflex in congenital nystagmus. Invest Ophthalmol Vis Sci 1985;26:1043-50. 5. Gresty MA, Halmagyi GM. Abnormal head movements. J Neurol Neurosurg Psychiatry 1979;42:705-14. 6. Anagnostou E, Spengos K, Anastasopoulos D. Single-plane compensatory phase shift of head and eye oscillations in infantile nystagmus syndrome. J Neurol Sci 2011;308:182-5. 7. Jan JE, Groenveld M, Connolly MB. Head shaking by visually impaired children: a voluntary neurovisual adaptation which can be confused with spasmus nutans. Dev Med Child Neurol 1990;32: 1061-6. 8. Dell’Osso LF, Daroff RB. Congenital nystagmus waveforms and foveation strategy. Doc Ophthalmol 1975;39:155-82. 9. Abadi RV, Dickinson CM. Waveform characteristics in congenital nystagmus. Doc Ophthalmol 1986;64:153-67. 10. Gresty M, Halmagyi GM, Leech J. The relationship between head and eye movement in congenital nystagmus with head shaking: objective recordings of a single case. Br J Ophthalmol 1978;62: 533-5.
