ml l 1 mum«msmww \‘HNIWIN “WWW lllllllllllllllllllllllllllllWilli L 293 10666 1592 LIBS-ZAEY {32122133} State . 523.3 3.: Sit)’ "Lg“ V‘— This is to certify that the dissertation entitled Visual, Vestibular. and Mechanical Factors in Chameleon Head Movement presented by Martha Flanders has been accepted towards fulfillment of the requirements for Ph.D. degree in Neuroscience/ ZOOIOgy Q E Ono Major protes or Date June 7, 1981:, .MSU is an Affirmative Action/Equal Opportunity Institution 0-12771 M‘- .--' __'-"-"—' _... MSU RETURNING MATERIALS: Place in book drop to ”“3155 remove this checkout from your record. FINES will be charged if book is returned after the date stamped below. $193588; OCT 0 3 2005 .WM-- 12 1 3 o 6 OCT 15 ’Ut§%?12u09 W31 8 2010 1 . ' 09 r330 1Q .0614 13 VISUAL, VESTIBULAR, AND MECHANICAL FACTORS IN CHANELEON HEAD MOVEMENT BY Martha Flanders A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Neuroscience Program and Zoology Department 1984 ABSTRACT VISUAL, VESTIBULAR, AND MECHANICAL FACTORS IN CHAMELEON HEAD MOVEMENT BY Martha Flanders In the African chameleon, a.moving cricket elicits a visually guided, pursuit head movement that aligns the tongue with the bait. Moving the chameleonfls body elicits a vestibularly mediated, stabilization head movement that keeps the head stationary in space. The independence of visual and vestibular head move- ments was tested by individually characterizing each, and comparing these characteristics to those of head movements made during combined bait and body movement. Results show a visual-vestibular interaction in which the timing of pursuit head movement is improved during combined bait and body movement. Mechanical analysis showed that equal amplitude and frequency, visual and vestibular movements of the head rela- tive to the body are mechanically similar, and suggested continuous neural control of head position. ACKNOWL EDGMENTS Thanks to Dr. James L. Edwards for financial support and laboratory space in the first three years of my graduate work. The research in Chapters 2, 3, and 4 was supervised by Drs. James L. Zacks (Psychology and Zoology), Erik Goodman (Electrical Engineering and Systems Science0,land Robert P. Hubbard (Biomechanics and Metallurgy, Mechanics and Materials Science). Dr. James H. Asher, Jr. and Walter E. Chapelle, PJL (my Daddy) helped with Appendices A and B, respectively. Thanks again to Daddy for making the apparatus, and to Hank Wieferich for assembling the mechanical parts. Special thanks to Dr. Hubbard for guidance. This research was supported by the Zoology Department and the Neuroscience Program, by grants from Sigma-Xi, and by my wonderful husband. ii TABLE OF CONTENTS L I ST OF TABLE S O O O O O O O O O O O O O O 0 LIST or FIGURES . . . . . . . . . . . . L . CHAPTER 1: CHAPTER 2: CHAPTER 3: CHAPTER 4: APPENDIX A: APPENDIX B: INTRODUCTION AND SUMMARY . . . . BINOCULAR HEAD TRACKING IN THE AFRICAN CHAMELEON O O O O O O O O O O 0 Introduction . Methods . . . . Results . . . . Discussion . . . VISUAL AND VESTIBULAR INFLUENCES IN CHAMELEON HEAD MOVEMENT: CHARACTERIZING NEURAL INTERACTION . . . . . . . Introduction . Methods . . . Results . . . Discussion . . Introduction . . Methods . . . . Results . . . . Discussion . . . STATISTICAL ANALYSIS . . . . . LISSAJOUS ANALYSIS . . . . . . BI BLIOGRAPHY I O O O O O O O O O O O O O O 0 iii _THE MECHANICS OF CHAMELEON HEAD MOVEMENT Page iv \ONDOlUl 01 47 59 62 Table Table Table Table LIST OF Pursuit Data . . Combined Data . Time Lags . . . Sinusoidal Motion iv TABLES Page . . . . . . . . . . . . l3 . . . . . . . . . . . . 29 . . . . . . . . . . . . 31 LIST OF FIGURES Page Figure 1: The African Chameleon in Nature. . . . . . 2 Figure 2: The Motion Producing Apparatus. . . . . . 7 Figure 3: Film Data. 0 O O C O O O O O O O O O O O O 10 Figure 4: Position vs. Time for Pursuit Head movement 0 O O O O O O O O O O O O O O O O l 1 Figure 5: The Pursuit Model. . . . . . . . . . . . . 14 Figure 6: Test of the Model. . . . . . . . . . . . . 17 Figure 7: Pursuit Head Movement with Body Stationary or Moved by Apparatus. . . . . 26 Figure 8: Frequency Response for Sinusoidal Head movement 0 O O O O O I O O O O I O O O O O 27 Figure 9: Kinematically Identical Visual and Vestibular Head Movements. . . . . . . . . 34 Figure 10: Deep Neck Musculature. . . . . . . . . . . 38 Figure 11: neCK ElastiCitYO I O O O O O O O O O O O O 41 Figure 12: Muscle Activity During Mechanically Similar Vestibular and Visual Head Movements. . . . . . . . . . . . . . . . . 43 Figure 13: EMG Data. 0 O C O O O O O O O O O O I O O 45 Figure 14: Test for Sinusoidal Motion. . . . . . . . 50 Figure 15: Linear Regressions of Time Lag vs. Frequency. . . . . . . . . . . . . . . . . 55 Figure 16: Gain and Phase Calculation. . . . . . . . 60 CHAPTER 1: INTRODUCTION AND SUMMARY The goal of this research was to find out what makes a lizard move its head the way it does. The African chameleon was an ideal subject for this motor control study, because of the conspicuous binocular head tracking exhibited by this lizard during its normal feeding behavior. Baby chameleons are born live and almost immediately begin track insects. Throughout life, the chameleon climbs on small branches as it aims both eyes and a projectile tongue at moving insects (Figure 1). Movement of the head relative to the body is guided by the vestibular system during body movement, guided by the visual system during insect movement, and always constrained by the mechanics of the head-neck system. The following chapters examine the relative contribution of each of these systems to coordinated head movement. Chapter 2 gives a qualitative description of feeding behavior in the African chameleon, and a quantitative char- acterization of the pursuit head movement system. Within a certain range of amplitudes and frequencies, a hungry chame- leon will follow a cricket, moved back and forth sinusoidal- ly, with a sinusoidal head movement and no apparent eye movement. The head lags behind the cricket and does not move as far in either direction as does the cricket. An Figure 1: The African Chaneleon.in Nature. The chameleon climbs on small branches as it hunts for insects. During locomotion or branch movement a vestibular reflex keeps the head stationary in space. A visually guided pursuit head movement follows a moving insect. Here, a female Chameleg sens: galensis aims both eyes and a projectile tongue at a short-horned grasshopper as it lands on a leaf. 3 equation is derived to predict gain (amplitude of head movement/amplitude of bait movement), given the chameleonfs time lag and the stimulus frequency. The most striking prediction of the equation, which is confirmed by the data, is that the head becomes aligned with the cricket just at the point when the head is momentarily stationary. Chapter 3 presents the central theme of the disserta— tion. The independence of visual and vestibular head move- ment systems is tested by individually characterizing each, and comparing these characteristics to those of head move- ments made during combined visual and vestibular stimula- tion. The visual (pursuit) head movement system was characterized in Chapter 2. Head movement controlled by the vestibular system was characterized by rotating the chame- leon's body back and forth sinusoidally, around the neck. Within ranges of amplitudes and frequencies the chameleons' vestibulo-collic reflexs (VCR) were observed to keep their heads stationery in space. The hypothesis that visual and vestibular head movement systems are independent predicts that during combined body and bait movement the VCR will respond to body movement by stabilizing the head, so that the characteristics of the movement of the head in space will be the samelaS'those of pure pursuit. When 10° bait movement was combined with 5° body movement, either in phase at the same frequency or at a lower frequency with no consistent phase relationship, the resulting movement of the head in space was sinusoidal, at 4 the same frequency as the bait movement, and had the time lag/gain relationship characteristic of pure pursuit. How- ever, in 17 of 18 trials, for a given chameleon, on a given day, time lag was shorter during body movement. Chapter 4 tests and supports the hypothesis that pur- suit and VCR.movements of the head relative to the body are mechanically similar and therefore place similar demands on the muscular system. Since there is no movement of the head in space during perfect VCR, visual and vestibular head movements are mechanically similar only if the headfls iner- tial resistance to acceleration is negligible. The compari- son of inertial and elastic forces for sinusoidal head movement shows that the neck is highly elastic and inertial forces are negligible. Electromyographic data show similar imuscle activity during pursuit and VCR movements, and sup- port the importance of elasticity in the chameleon head-neck system. CHAPTER 2: BINOCULAR HEAD TRACKING IN THE AFRICAN CHAMELEON Introduction The African chameleon has a unique feeding strategy. These arboreal lizards capture insects by shooting a long sticky tongue straight out of their mouths, off a special- ized hyoid apparatus (Murphy, 1940L. The chameleon scans its environment with large amplitude, independent, saccadic eye movements (Walls, 1942; Mates, 1978). When an insect is spotted, the head is aligned, both eyes come forward to fixate the target, and the hyoid is extended in what is known as the ”initial protrusion! (Gans, 1967; Bellairs, 1970). During initial protrusion the chameleon decides how far to shoot its tongueeby accommodation (focusing) rather than triangulation (as one might expect) (Harkness, 1977). The chameleon has a deep convexiclivate fovea that acts as a focus indicator (Harkness and Bennet-Clark, 1978) and may work along with short depth of field from iris dilation, and efference monitoring of motor commands to lens muscles, to help the chameleon judge distance. When an insect moves past the chameleon during initial protrusion and binocular fixation, the chameleon moves its head, and not its eyes, in smooth pursuit. Foveation is a necessary part of pursuit, but an equally important goal is 6 to align the tongue with the prey. Unlike mammals and fish, who pursue with visual fixation followed by counter-rotation of the eyes during head movement (Bizzi, et al., 1971: Lanchester and Mark, 1975), the chameleon often fixes its eyes in its head and pursues by performing head movement alone. This chapter reports that within a certain range of amplitudes and frequencies, hungry chameleons will perform sinusoidal pursuit head movement, with no apparent eye move- ment. In spite of individual timing differences, and over their entire range of frequencies, all of the chameleons tested tracked sinusoidally moving crickets using a strategy that put the head at zero velocity when the bait was direct- ly in front of it, and minimized retinal slip. Methods {Five female Chameleg senegalensia, weighing between 20 and 60 gms., were obtained commercially and housed in a room with high humidity, 12 hr./day florescent lighting, and additional incandescent lights turned on as needed for bask- ing and to adjust the temperature. Preferred body tempera- ture was about 80-85°F (27-29°C) in the day, and 70°F (21°C) at night. The chameleons ate about six crickets/day. The crickets were fed a high calcium diet (Allen, personal communication). The chameleons were pretrained to feed on a bait-moving apparatus (Figure 2). The chameleons clung to a stationary M Figure 2: Ll; l 1U The Motion Producing Apparatus. With her tongue in initial protrusion, the chameleon clings to a movable perch and aims at a movable cricket. The forward position of the perch can be adjusted so that the angular, horizontal rotations of the cricket and/or perch (moved by gears and motors) are centered at the chameleon's neck. (Apparatus by ChapelleJ 8 perch and aimed eyes and tongues at moveable crickets (the bait and reward for pursuit). The bait was moved by a servomotor driven by a function generator chip. Angular, horizontal, sinusoidal motion between 0.5 and 4.0 Hz was produced with amplitudes up to about 20°. As a chameleon grasped the perch, its position could be adjusted so that the axes of bait movement and head movement were concentric. Marks were painted on each animal's head and eye lids with black acrylic, to aid in position measurement. The motions of the bait and head were recorded on 16 mm film at 50 f.p.s., with a telephoto/macro lens on a Bolex camera mounted about 2 meters above the apparatus. Ambient temperature was not controlled during film exposure but was usually near 85°F (29°C) due to incandescent spot lighting. A Lafayette stop motion projector was used to view the film. Pursuit sequences were projected onto an angular grid with the grid center at the center of the bait and head rota- tions. Generally the chameleons grasped the perch firmly and moved only their heads, but when small body movements occurred, the grid was adjusted frame by frame to keep the coordinate system centered at the chameleon's neck. Error introduced by moving the coordinate system was less than the error of position measurement off the grid (about i INF“. Film sequences containing at least three complete cycles of sinusoidal, steady state, pursuit head movement without eye movements were analyzed. A graphical technique was used to calculate the average amplitude gain of the head 9 movement relative to the bait movement and the phase lag of the head movement (Appendix B). Results On film the motion data appeared as shown by the series of photographs in Figure 3A. Figure BB places the photo- graphs in one complete movement cycle, sampled at 1/4 sec. time intervals. The bait moves back and forth sinusoidally and the head follows in pursuit. Because chameleons do not move their eyes within the lids, the lines painted on the eye lids indicate that the eyes were fixed in the head during this cycle. Before initial tongue protrusion and during less vigorous pursuit, the chameleons sometimes locked one eye straight ahead while the other continued to scan the environment or moved in smooth or saccadic pursuit. Small conjugate saccadic movements were also observed. The behavior of the eyes was closely observed during each film exposure and measurements were taken only from cycles judged to be free of eye movement. Figure 4 shows a representative pursuit sequence, as measured on the angular grid. Bait position (closed cir- cles) and head position (open circles) are plotted frame by frame against time for five continuous cycles. Despite the error in measurement, pursuit head movements always appeared to be smooth and the data were well fitted by sine waves of the same frequency as the bait movement (Appendix A). In the frequency range from 1.0 to 2.8 Hz., the chameleons .Aocfia canny moon was Accua xofinuv Damn can no mcofiuoe uanmcm msoscaucoo may :a Amid ambasscv smoumouonm some no :oHumooH onu moumEaxOLQQm Amy smoum one .ucoeousmsoe :oHuumom cm was memo was new: on» so command mocfifl one .w.o usonm mo :fimm a can ..oomE and no mnH mean a spa: ..m= hon.o um ucoEo>OE coo: ufismusm no oHoao oco msuouuom Adv :ooHoEono on» .mHm>uou:fi .oom «\H as o>onn Ecuu monmmumouosm .sumn Baum um ousmfih ‘(viifV' ,v‘r 11 .uann on» use use magnum on osmcou no: use uoam cooaosmno on» uwsmusm no mofiumo Hauo>om momma .coHuHmom menu on» us one one: can name cogs Dawsoommmmamcxmom an m“ one: can awn» ovauoz. .um ¢.a um unm5o>oe one: ufismuzm n mcqusn mHm>uouca Abacum ocov .ooms on on czogm ma on one: can 3 Damn mo cofiufimom unasmcc .usoflobo: use: unsmusm mom alum. .99 scuuumom um Aim}: 2: 0000 as. 0000 «E 8hN «E OOON - a J 4 1.104.. T d d a q u .d d d d d a . .0, o oo coo-coo 00.0 o o «986 o o 1 n3. 0 o 0 n6 0 o o 1 5 O 8 pi p 0 o 0" fl < o O 3:... o o oo oo o o o h. WK“. Do o 000 o oo o 1 m 00 o. com owe. .800 0.000008 as L mm mm o- 0 «000000- 0000. o 1 l 0 one: so 0 L 02% .0— 238.»— 2:08. 2:8” 1 d ld d T q d d d d d d d d H d d d 0°— 00 can... coo-so .. 00... 00000 no. ooowomooooo so ._ m 0 000 I 1 J. o o .XEr o o O o 30 o l o C O D o C O o O .- llomlu. oo o o x no 00 onl o 9. nv.0 O o GOOOOOQOO o 000 o O o 00 O 00.. H . . O . o I- 99600 0 Ac 0 OQXbo o 9X€m#do n m o o 00 o o s L 1 so. so 0000 so a 0 so I sees as... o o .0. omsmfim 12 performed sinusoidal pursuit with a negative phase (or time lag, (kt), and an amplitude gain (the ratio between peak head displacement and peak bait displacement, H/B) less than one. Table 1 shows measured time lag and gain for the pursuit movements of five different chameleons duing 10° amplitude bait movement, on several different dates. The time lag of an individual chameleon varied little relative toldifferences between animals, and was not significantly correlated with frequency. The chameleons moved with large gains (near 1‘0) at low frequencies, and with smaller gains at high frequencies. The position vs. time record in Figure 4 shows that although the chameleon tracks the bait with a time lag, she uses an amplitude that puts the head at peak displacement (and zero velocity) when the bait and the head are at the same position. This phenomenon was observed in all five chameleons, even though time lags varied from 52J3(Susan) to 144.4 (Bette) msec. and frequencies ranged from 1.0 to 2.8 Hz. Figure 5 shows the observed pursuit strategy in the form of a generalized position vs. time graph. The bait movement (thick line) is followed by a pursuit head movement (thin line). B is the amplitude of the bait movement (about 10°). The head is at peak displacement when the bait and head are at the same position, so that head amplitude (H) depends on the time lag (23t) and the period of a cycle. Time lag varies between animals and between trials, and the 13 Table l: Pursuit Data. Characteristics of pursuit head movement with body stationary. f At Measured Model Date Chameleon (Hz .) (msec .) H/ B H/ B A H/ B 1/02 Susan I 2.8 89.3 0.25 0.21 +0.04 I 2.0 52.8 0.82 0.79 +0.03 I 1.4 59.5 0.86 0.86 0.00 2/08 I 2.8 71.4 0.38 0.31 +0.07 I 1.4 55.6 0.82 0.88 -0.06 m 2.0 54.2 0.65 0.78 -0.13 I 1.7 66.1 0.72 0.76 -0004 2/27 I 1.4 61.5 0.88 0.85 +0.03 I 1.2 63.0 0.91 0.89 +0.02 2/08 Carol X 1.0 108.9 0.81 0.78 +0.03 7/02 X 1.0 97.8 0.80 0.81 -0.01 2/08 Bette #- 1.0 144.4 0.66 0.61 +0.05 7/01 Hilde A 1.0 90.6 0.90 0.84 +0.06 7/08 A 1.7 103.7 0.55 0.45 +0.10 7/03 Lisa . 1.0 86.1 0.90 0.85 +0.05 7/04 I 1.4 93.6 0.70 0.67 +0.03 0 1.0 76.1 0.77 0.88 -0.11 7/08 I 1.2 72.4 0.88 0.85 +0.03 0 1.7 73.4 0.79 0.70 +0.09 7/13 I 1.4 80.9 0.72 0.75 -0.03 7/17 I 2.0 99.7 0.44 0.31 +0.13 14 B 0‘ H At 1 L L t Figure 5: The Pursuit Model. Angular position vs. time for sinusoidal bait movement (thick line) and pursuit head movement (thin line) to the left (down) and right (up) of a zero position. Time lag (At) is shown at the zero crossing and again between the peak displacements. Amplitude gain (8/8) is predicted using the formula for point B (head amplitude) on the line with ampli- tude B (bait amplitude). 15 period of a cycle depends on the frequency set with the apparatus. In order to quantitatively test the usage of this strategy, an equation was derived to predict amplitude gain (H/B), given time lag (rat) and period (T) or frequency (fal/T). The equation for predicted gain: 11/8 8 sin 2'Tlf[(1/4)T + At] is derived from the angular position equation for sinusoidal motion: 8 a A sin 2TTf(t), where 9 is angular position, A is peak displacement, and t is time. For position H, on the line with amplitude B, H = B sin 21Tf[(l/4)T + At]: where (1/4)T is the time at point B and point H follows B by (it. Use of the head amplitude predicted by the model (rather than a larger or smaller head amplitude) minimizes the amplitude of the sinusoidal motion of the bait relative to the head (retinal slip for pursuit head movement without eye movement). This can be shown by representing a time lag - frequency combination as a phase angle: ‘P a 21mm: The motions of the bait (B) and head (H) can be represented as two vectors separated by the phase angle‘P. Since'fi is a A vector, it can be separated into two components: Bx is —3 parallel to H, and By is orthogonal to H. The amplitude of the difference between the two sine waves (relative motion 16 and retinal slip): 3-3 a (Bx-H) +13; is minimum when H is equal to Bx’ This occurs when H . B cos‘P a s sin (90° HP) =- 8 sin 2’Nf[(1/4)T + At] as shown in the model. Time lags of the five chameleons performing at frequen- cies between 1.0 and 2.8 Hz were used to calculate the gain predicted by the pursuit model equation for each pursuit sequence. The difference between measured gain and pre- dicted (model) gain is shown in the last column of Table l. The average difference between measured gain and model gain (AH/B) was 0.018 which represents about a 0.18° overshoot of the predicted head amplitude. Using a t-test, this difference was not found to be significantly different from zero (0‘- .05). (This test was also performed using an alternate measurement of time lag and the head undershot the model by 0.1l°. See Appendix A.) Figure 6 shows how well the data fit the model. Observed gains for each trial are plotted against the gain predicted for that trial, using time lag and frequency. Observed and predicted values have a significant positive correlation (correlation coefficient = (L95L. The ye intercept of the regression line, however, is significantly different from 0.00 (Snedecor and Cochran, 1971). This 1.00 - 0.90 0.80 0. 70 ’s 0. 60 0.50 Gain 0.40 0.30 0.20 0.10 - 0.00 Figure 6: 17 l ’1 l 1 l l l l i l 0.20 0.40 0.60 0.80 1.00 1 Gain = sin 211'“ I T + At) Test of the Model. Observed gain (H/B) vs. gain predicted by the model equation. Symbols represent trials shown in Table 2. A linear regression line is shown. Correlation coeffi- cient - 0.95. 18 shows that low amplitude head movements overshot the predicted head amplitudes. Head movements observed during 5° bait movement were also greater than predicted by using the model equation with B 8 5°. Susan and Lisa were tested with 5° bait movement, at 2.8 and 2.4 Hz respectively. In each case the chame- leon's time lag was within her normal range (58.5 msec for Susan and 78.2 msec for Lisa), but the head overshot the amplitude predicted by the model by about l.2°. The range of frequencies and amplitudes used was limited by the chameleons' performances. At low amplitudes and low frequencies the chameleons struck and ate the cric- kets without performing three cycles of pursuit. At high amplitudes and frequencies the chameleons showed no head movement at all. For 10° bait movement, the frequency range of a chameleon was limited by her time lag. When a frequency-time lag combination represented a 90° phase lag, the model predicted a 0° head movement. For example, Bette, with her 144.4 msec. time lag, approached a 90° phase lag at 1.7 Hz. Thus it was impossible for the slower animals to perform pursuit head movements at high frequencies. Figure 4 shows that after several cycles of pursuit the chameleon shot out her tongue to strike and eat the bait. The chameleons were most likely to hit the bait if they struck near peak head displacement, where the bait was in front of the head and moving slowly; The strike shown in Figure 4 was successful (a hit) for this reason, but the 19 chameleons did not always choose the optimal time to strike, and often missed. For the trials in which the strike was recorded on film, the chameleons missed 69% of the time. Although the head was usually stationary when the bait was directly in front of it, it took about 40 msec. for the tongue, once fired, to travel its distance (and some addi- tional time for motor commands to reach tongue muscles), so that by the time the tongue reached its target the bait had often passed. The fact that the pursuit system does not appear to take this additional time lag into account, and the variability in head position at the time of strike, suggest that the pursuit movement and the decision to strike are two separate processes. Discussion The study of pursuit head movement made in response to horizontal, sinusoidal bait movement, demonstrated that since head movement cannot track bait perfectly (due to a time lag), the chameleon uses a strategy that brings the head in line with the bait when the head is at zero velo— icity. Since the tongue must be shot straight out of the mouth, this might seem to be an ideal feeding strategy. But the chameleons often missed. (They very rarely miss sta- tionary targetsJ The head amplitude represented in the model minimizes retinal slip velocity and acceleration. Thus, the goal of the pursuit head movement system may be to foveate the prey and judge its distance, rather than to eat it. 20 Several lines of evidence suggest that the chameleonfis pursuit head movement is visually guided by a continuous position tracking system. Interupting vision abruptly stops the head movement. The chameleon head-neck system is highly elastic (Chapter 4), and head movements did not (to the resolution of my measurement) contain free-swinging (or ballistic) segments. In nature the chameleon rarely encoun- ters a cricket moving sinusoidally at a single frequency, but the experimental use of sinusoidal motion may have revealed a continuous position tracking system basic to pursuit. Other authors have emphasized prediction in pursuit movement. Human smooth-pursuit eye movements follow target motions of low frequency with negligible phase lag due to sine wave prediction (Lisberger, et al., 1981). If the chameleons predicted the wave form of the bait movement rather than tracking it continuously, they should have im- proved their time lags from cycle to cycle (or from trial to trial). This was not seen. Lanchester and Mark (1975) argued against predictive pursuit for fish swimming toward falling bait, by modeling the path of the fish as one in which the mouth always points directly toward the bait. That study did not take into account a time delay between sensing the position of the bait and bringing the body into line with it. If fish have time lags between bait and head movement similar to those seen in chameleons, the movements that fit those authors' model were, in fact, predictive. 21 Further studies of chameleon head movement have shown that chameleons also use the pursuit strategy described above, for pursuit during the body movement. During passive body rotation, a chameleonls individual time lag was always shortened (Chapter 3) for higher gain, less retinal slip, and more successful strikes (improved pursuit performance). CHAPTER 3: VISUAL AND VESTIBULAR INFLUENCES IN CHAMELEON HEAD MOVEMENT: CHARACTER- IZING NEURAL INTERACTION Introduction In vertebrates, head and eye movements are guided by visual, vestibular and perhaps proprioceptive stimuli. A small, smoothly moving visual target is followed using pur- suit head or eye movement s’ystems. Movement of the entire visual field, relative to the eyes, is minimized using visually guided, compensatory (optokinetic) movements. Dur- ing body or head movement, receptors of the vestibular system (utricle and semicircular canals) also mediate com- pensatory movements that work to stabilize the head (vestibulo-collic reflex, VCR) or eyes :(vestibulo-ocular reflex, VOR) in space. Interactions between visual and vestibular movement systems have been studied extensively (Henn, et al., 1980) in the hope that they will reveal mechanisms basic to integrated sensory-motor function. The fundamental problem in studying visual-vestibular interaction is recognizing visual and vestibular components of a movement guided by both stimuli, One approach is to individually characterize the visually guided movement and the vestibular, stabilization.movementq and compare these characters to those of movements made during combined visual 22 23 and vestibular stimulation. This tests the hypothesis that movements during combined stimulation are the sum of normal visual and vestibular movements. Rejection of this hypothe- sis is evidence for interaction between the two systems, while failure to reject supports their independence. Investigations of smooth pursuit eye movement and VCR in primates, have been guided by the idea that VOR may be suppressed during head tracking to improve pursuit perform- ance (Robinson, 1977). Lisberger, et a1. (1981) character- ized the pursuit system in monkeys, animals with nearly perfect VORs, and then combined oppositely directed visual and vestibular stimulation by rotating a visual target exactly with the animals' heads. The result was failure to reject the independence hypothesis. .However, when visual and vestibular movements are exactly opposite, an infinite number of compromising interactions (vestibular suppression along with a change in pursuit) could give this same result. Other investigators have rejected the independence of visual and vestibular movements. Bock (1982) tested human smooth pursuit and VOR in a conflict situation where visual and vestibular eye movements were in opposite directions, and concluded that eye movements were sometimes predominant- ly visual and sometimes predominantly vestibular (a switch- ing type, non-linear interaction). In geckos, the combina- tion of visual (optokinetic) and vestibular (vestibulo- collic) stimuli in the same direction was found to improve 24 the performance of one or the other or both systems (Meyer, et al., 1979). Whether the visual-vestibular independence hypothesis is rejected or not, the with-or-without-body-movement para- digm does not always allow quantification of visual and vestibular components of a movement made during combined stimulation. The African chameleon, with reliable pursuit head movement characteristics, and a nearly perfect vestibulo-collic reflex (VCR), has exhibited separately recognizable visual and vestibular components of complex movement. Methods Data were taken from five female Chameleg_sgnegalensis, obtained commercially and pretrained to feed on the testing apparatus (Figure 2). The apparatus allowed separate con- trol of visual (cricket movement) and vestibular (body move- ment) stimuli. The chameleon's neck was centered over the axis of rotation of the apparatus as the cricket and/or the animal's body was moved back and forth sinusoidally at various frequencies (0.7 - 2.8 Hz) and amplitudes (10° cricket movement and.5° body movement), in the horizontal plane. Resulting head movements were filmed from above (16 mm, 50 f.p.s.). Films were projected onto an angular grid to measure cricket and head positions with an accuracy of about i,1°. 25 Results A representative pursuit performance is shown in Figure 7A. Angular positions of the cricket (closed circles) and the head (open circles) are plotted against time as the bait is moved 10° to the right and left of a zero (straight forward from the stationary perch) position, and the head follows in a motion not significantly different from sinu- soidal (Appendix A), with no apparent movement of the eyes relative to the head (Chapter 2). For a 10° amplitude cricket movement, the frequency response of one chameleonfis pursuit system (data from Lisa, Table l) is shown in Figure 8 (closed circles). In Chapter 2 an equation was derived to predict gain (amplitude of head movement/amplitude of crick- et movement, H/B) given time lag and frequency. The solid curved lines in Figure 8 show the gains that correspond with constant time lags according to this equation. Lisa per- forms pure pursuit with a time lag of around 80 msec. VCR performance was measured by rotating the chamele- ons' perch sinusoidally around the neck with an amplitude of 5° and no visual target. Vestibular reflexes are usually tested in the dark to exclude visual interactions, but because of the plasticity of vestibular function (Barr, et al., 1976) it is preferable to test VCR under conditions comparable to those used for combined visual and vestibular stimulation. The chameleon, well known for her independent, saccadic eye movements (Mates, 1978), excludes the possi- bility of visual interactions by failing to suppress $00. 26 .2256... .68. .um 2:..e 53 confinEoo acme—okra... Dawn .5. o4 E can Jame—mega “Son .5. arm 2 9.2.6 mao>uousa “oeouu 03“: .oomE 3. us cacao ma woman a“ .3 coo: use .3 mean no ceauamom unusmcc .msuouommc an echo: no hue-5.333 moon nun: analogy: coon awn-mum.— 2. 353m GODQEV ”2.5. com— 003 Don— OON— Do: 000— oom 08 00h 08 com Gov Dom CON cow 0 q I. I q I q I I I I I I I I I I I 0 Q m ” Or 0 o o 0 O O O C O O H e o o o e o o o 1 fl 0 O. O O O O O 1 I. O P O r a n 1 1 00 O O O O L e o o H O O O n m o L H o o o o l o O o a O o 1 o m o I o e o e 0 loop A.OONEV mzrp com. 00: Comp 009 CO: ooh: com 00m 005 com com 00v 000 DON cow 0 I I I I I I I I I I I I I I I I I O m e e e e e e .. 0 e o I e - o o o o o o o o «2.. O O O O O O O O H... o o e o e o 1 o t o o oo o o o o H O O O O O O H o o o o 1 F 0.. .0 0° 3 A< 7 2 . .mmoH mam» omega mean: Homes oasmusm on» an couoficoum madam on» can .cmoum mesons loom .m> omega emu no mood use» unconsoo one momma co>uso was .ucoEo>oe been new moon Gaga—=03 soon one 23:92:: uno5o>oe “Zoo .32: ucoeo>oe moon mcausm aco6o>oe coon Homeomsnfim mom csonm one nuaumauouoouosu woman can Emu dead; 25 mo uses—obs: coon Hogan—Ea uou uncommon moss—503m “a 3:3...— ma. 42.: 6:03am: finzv 6533: mg“ aim cg. c._ m.u 0A“ v; o; 1 I ll. J be :35... o .0368 c: on. _em omega .0 So 359800 n. “op amp ov— om .05 omega c. 35950 O I E . 6er 8 c: on. I co co. om. cc. 0 on. u on on m on . om 20> x x x x x .0 cu) m x .- «7% On. mu. (B/H) Ute!) The L1 28 saccades during compensatory head movement, as would be ex- pected for optokinetic stabilization (Pratt, 1982). The chameleons were observed to keep their heads stationary in space for several consecutive cycles of body movement, at each frequency. This excellent vestibular performance (shown in Figure 8 as a gain of 1.0 and a phase lag of 0°) was not surprising since the amplitude and frequencies used were similar to those produced by the chameleonls normal locomotion (measured on film in this laboratory). According to the independence hypothesis, the combina- tion of cricket movement with body movement (in the range of excellent VCR) will produce a head movement (in space) identical to that of pure pursuit. When 10° amplitude bait movement was combined with 5° amplitude body movement, either in phase at the same frequency, or at a lower fre- quency with no consistent phase relationship (as in Figure 7B) the resulting movements of the head in space were sinu- soidal at the same frequency as the cricket movement (Appen- dix A) and amplitudes corresponded with time lags as pre- dicted by the pursuit equation. For the 23 combined cricket and body movement trials shown in Table 2, predicted and observed head amplitudes had a difference not significantly different than 0.00 (t-test,oe ”Snob msonm we: .32» 3333 .m .2332. been 26.3 we: 3.35 38> 2 .ocaa cams on» an csonm moon on» on o>fiuoaou moon on» no usoao>os ecu mcfiusc moaomse goo: puma can unmau no can .musolobot coon donut, use Housman—mob Honwlwm mudogsaoo: menu—5 muwmuuum muons: «NH 353... _ _ . _ . _ comp _ ” WW 5 \V/ \V/ \V/ a / > > )- a i x Q Q Q Q < < . SENT. P r ii I V ._r i .fimr the- m. 44 Figure 12B shows a mechanically similar pursuit head movement (thin line) following an apparatus produced bait movement (thick line). Since there is negligible inertial resistance to pursuit head movement, the activity centered around peak displacement, again, counteracts elastic resist- ance to displacement. Although there is less movement arti- fact in the pursuit EMG belectrode leads were taped to the body), the actual muscle signal appears to be very similar to that recorded for VCR. Figure 13A and B compare the muscle signals for VCR (A) and pursuit (B) movements at a higher graph speed, along with the actual polygraph traces from the apparatus (the thick lines in Figure 12L. Figure 13C shows several cycles of VCR during which the chameleon had both eyes closed (excluding the possibility of a visual contribution to head stabilization). Discussien Results of all three experiments suggest that kinema- tically'similar visual and vestibular movements are also kinetically similar and mediated by similar muscle contrac- tions. Passive body rotation of anesthetized and dead ani- mals showed that elasticity was greater than inertia. Quantification of these parameters was in agreement with this result. EMG patterns were similar for VCR and pursuit movements and again showed the importance of elastic resist- ance to displacement. A) 3 _.L..|._ _ Al—L - ' fif'r r-vwnfi' fir, er-v C) 45 a) i WWW Figure 13: EMG Data. Records from Figure 12 run at a higher graph speed to-show that although there is more movement of the base line during VCR (A), the high frequency muscle signal is similar to that of pursuit (B). Curved lines show angular position of the body (A and C) and bait (B) across time. During the seven cycles shown in C, the chameleon kept her head stationary in space with both eyes closed. Scale is shown in Figure 12. 46 Viscous resistance to velocity was not measured in this system. Moment due to viscosity is zero at peak displace- ment.of a sinusoidal movement (since velocity'of the head relative to the body is zero), and viscosity, therefore, was unimportant in the comparison of inertial to elastic forces. Although the EMG analysis was limited, the results do not show much activity during peak velocity, as would be expect- ed if viscosity were a significant resistive force. Fric- tion in the neck joint also was not measured. In a study of monkey head-neck mechanics, friction was measured and was found to be negligible (Bizzi, et al., 1978). In the chameleon, head stabilization during body move- ment is completely due to active muscle contraction. Furthermore, since there is no passive movement of head relative to body within normal acceleration limits, it is unlikely that the stretch reflex contributes to stabiliza- tion. Proprioceptive information from the actively moved head should be the same in VCR and pursuit situations. The importance of neck elasticity in this system is consistent with the spring model for motor control (Bern- stein, 1967: Bizzi, et al., 1976). According to this theory, head (or limb) position is controlled by adjusting the stiffness (or resting length) of antagonistic muscles. On the other hand, inertia and viscosity are not negligible at all joints. Other theories of motor control stress the use of momentum in movement coordination (Greene, 1982; Goodman and Kelso, 1983). APPENDICES APPENDIX A STATISTICAL ANALYSIS Statistical analysis was used to address.a number of questions underlying the assumptions and conclusions re- ported in the preceding text: 1. Is the head movement sinusoidal? Use of the equation: H/B . sin 2'n£[(1/4)T + At] to predict gain (Chapter 2), was based on the assumption that both the bait movement and the head movement were sinusoidal. The movement of the bait was apparatus con- trolled and had been checked on an oscilloscope by cross plotting it with a known sinusoid to assure that it was sinusoidal. For each set of data the bait and head posi- tions were graphically cross plotted for points in time (Appendix B) and were used only if they could be fitted (by eye)*with an ellipse. This Lissajous method gave a qualita- tive check to make sure all movements were roughly sinu- soidal. Five out of 49 trials that were analyzed, were not used because they were not fitted by an ellipse. Upon review of the film it was discovered that during these 5 trials the cricket was not well clipped to the apparatus and was vibrating as it was moved. 47 48 To quantitatively test the assumption that all head movements were sinusoidal, the Lissajous graphs were re- viewed to select the worst (or least well fitted by an e1- 1ipse) performance of each chameleon. For each of these trials head position vs. time data were fitted with a sine wave generated by the following formula: (9 a H sin 2Tif(t) where 9 is angular position (in degrees), H is the amplitude of the head movement (in degrees), f is frequency (in Hz), and t is time (reported in intervals adjusted from l/21 sec. to 1/24 sec. to match the film speed of a given trial). Carol's worst head movement occurred on 7/02 (Table 1) This trial was fitted with a sine wave of the formula: 9 =- 8 sin 2’” (1) (1/22) by using a t-test (Steel and Torric, 1980) to test the hypothesis that the difference between measured head posi- tions and those generated by the equation was 0.00° (Ho: 5 = 0.00°, H1: 5 ;‘ 0.00°). Table 4 shows position points gener- ated by using this formula, reported at 1/22 sec. (two frame) intervals. Figure 13 shows this rectified sine wave plotted on an angular position vs. time graph (closed cir- cles). Actual data from Carol's pursuit head movement is shown in the second column of Table 4. The rectified head movement is plotted in Figure 14 (Open circles). For each point in time head position was subtracted from the position on the true sine wave (third column, Table 4) and these deviations were averaged. The average difference between the sine wave and the head movement (5) in this case was 49 Table 4: Sinusoidal Motion. A sine wave generated using the equation for sinusoidal motion is compared to a representative pursuit head movement. The absolute values of angular positions (l8°|) are shown at 45 msec. (two frame) intervals. |9°| l9°l l6°l sine wave sine wave head - I6°l head MOU'IUIUUIOOUIUIUU‘UIU‘IOOUIOUIUIUI I HOOOOOOOOHHOOOOOOOOOO thQNQNQhNONhGQNQQObN e o e e e o “NOWWUUOUUOUUOWD‘DUOUN wmmslsldthwHH-mmmmslqmui-I ooeeeeeeooeoo'eoeeeeee NdUINfi-buonwmmNMin-ahummm 50 .AOV mucous sojwmom coo: confines.» s can Rev o>o3 on: we: iauoou o coosuon oocououuwc on» some on mean uncaomo couuoam one s canon. scum muons soquom .sofi—o: McQueen—sum you soon. a: 95m: muss: 3. 2. av 3333 an an as Va «a on 2 2 ST 9 o o v « Mludfidfidudq IdidquI‘Iqq L. 00 O LN In 0 o o L era. .‘ 'o O .m 0 Wk 10 O O O I .k r00 .0 51 -0.ll°, leading to a failure to reject Ho:d = 0.00 (n = 21, SE a 0.15, t a 0.73). Susan's performance on 2/08 at 2.8 Hz. (Table 2) was also given a t-test to see if the head's deviation from sinusoidal motion was significantly different from OJNFK In this case the head movement was fitted by the formula: 8 -- 5.5 sin 2'" (2.8) (1/24) with 18 data points (5 =- 0.49°, 55 = 0.39°) the calculated t value (t a 1.25) was less than the critical value (t(.05)15 a 2.131) leading to a failure to reject the hypothesis that the head movement was sinusoidal. Hilde's performance on 7/01 (Table 2) was fitted using the formula: (9 = 7 sin 2’Tf(l) (1/21) The hypothesis that d =- 0.00° was, once again, not rejected at the «X a .05 level of significance (n = 18, E = -0.33°, 35 = 0.21°, t = 1.56). Lisa's head movement on 7/03 (Table 3) was found to be significantly different from a sine wave of the formula: 9 = 9 sin 271(1) (1/22) (n a 18, E = 0.540, s; =- 0.22°, t a 2.45) at the °<= .05 level, although not at the OK = .02 level (t(0.05)15 = 2.131, t(0.02)15 a 2.602). The one set of data taken from Bette was extremely wel l fitted by the Lissajous ellipse, so it was not necessary to perform a t-test in this case. 52 Failure to reject the hypothesis that all but one of the worst head movements were sinusoidal supports the assumption that all head movements were sinusoidal. 2. Are the estimates of time lag obtained using a second method of measurement, significantly different than those reported in the preceding text? Time lags reported in the preceding text and used to test the pursuit model (Chapter 2), and to support the conclusion of decreased time lag during body movement (Chap- ter 3), were calculated by the Lissajous method (Appendix E». This analysis used all of the data points for bait and head movement in a given trial, to give an estimate of phase lag, which was converted to time lag using frequency. A second estimate for time lag in each trial was ob- tained by calculating the average number of frames between the zero position crossings of bait and head movements. Since a frame lasted about 22 msec. and time lags were usually less than 100 msec., this method gave a rather rough estimate. For the pure pursuit data (Table l) the average time between pairs of zero crossings had a standard devia— tion that ranged between i 0.0 and i 22.4 msec. The estimates of time lag calculated by the two methods were found to be significantly different from each other at the C5: .05 level (although not at the<9<= .02 level). A t— test was used to test the hypothesis that the.average~dif- ference between the two groups of measurements was 0.00 msec. (Ho: 5 =- 0.00 msec., H1: 5 :4 0.00 msec.). For pursuit 53 data, the average difference (5) was 6.21 msec. (n = 21, 85' a 1.94 msec.) leading to a t value of 3.21*. For combined stimulation data, the two groups of measurement were again significantly different (n = 23, d a 4.17 msec., s5 = 1.54 msec., t - 2.71*). These differences in msec. measurements represent an error of about 7%. In 34 of the 44 trials the time lag calculated through Lissajous analysis was larger than that measured from zero crossings. This difference probably represents a systematic error: ellipses could have been drawn too round, or the number of frames between zero crossings counted too conserv-l atively, or both. Notice that although the two groups of measurements are significantly different, the average dif- ference (6.21 msec or 4.17 msec, above) is less than the average error of zero crossing measurements (i 9.6 msecJ. In spite of the significant difference between methods of measuring time lag, either estimate can be used in the pursuit model (Chapter 2) to predict the amplitude of head movement. Use of Lissajous data predicts a head movement with a slightly larger amplitude than that observed (n = 21, d a 0.18°, 35 a 0.14°), while use of zero crossing data predicts a head movement of a slightly lower amplitude (n a 21, d a -0.1l°, 33' = 0.170). In both cases the difference between predicted and observed head amplitudes was not sig- nificantly different from d = 0.0° (for the Lissajous esti- mate of zAt, t a 1.26; for the zero crossing estimate of At, t = 0.65, 05:- .05). These convergent results support the validity of the model. 54 In Chapter 3 (Table 3), Lissajous time lags were com- pared for pursuit with and without body movement to show that the difference between the two (n =- 18, d’ =- 13.75 msec., s5 =- 2.17 msec.) was significant (t a 5.26"). This treatment difference is more than twice the difference be- tween the two measurement systems, and furthermore, compari- son of zero crossing time lags for the same 18 trials, shows the same trend (5 - 11.77 msec., s5 8 2.65 msec., t = 4.43**). In Figure 15, Lissajous time lags (top) and zero cross- ing time lags (bottom) are plotted against frequency for three different chameleons with (Open symbols), and without (closed symbols) body movement. The variance of the zero 2 crossing data (251.4 msec. with body movement, and 347.8 msec.2 without) is larger than the variance of the Lissajous data (221.8 msec.2 with body movement, and 242.0 msec.2 without). The regression lines show that in each case, time lags are decreased during body movement. The zero, or negative slopes of the regression lines are an artifact of the fact that only the faster animals will perform at higher frequencies. Regression lines for individual animals have positive slopes, and only in the case of Lisafls performance during body movement (the Open circles, phase or zero cross- ing) is the time lag data significantly correlated with frequency (correlation coefficient = 0.69,“= 0.05). Although the difference between Lissajous and zero crossing measurements of time lag for each trial is signif- icantly different from-d”= 0.0 msec., the extra measurement Figure 15: 120 I no - . A 100- I A I 90- ‘ . I__ At 70 r- O“~=-- 5— ' (msec.) Q -I‘UI-~~-__ 60- 6 I '““‘-—--- I I E) I C.) 40 - D 30 i- L L l l l ’1 1.0 1.5 2.0 2.5 3.0 trequency (Hz.) 120[. no . A 100 '- A A . w I- , .. a o B so - I A! _ Q j C I (msec.) A - O I 60- 0"“ “a ------------------- C) O (3 D 50 - O D I d D II ‘3 D 40 ' D D 30 D b |_ 1 1 1 J 1.0 15 2.0 25 30 frequency ( Hz. ) Linear Regressions of Time Lag vs. Frequency. Time lags for three chameleons were measured by A) phase and B) zero crossings. Symbols repre- sent trials shown in Tables 1 and 2. Solid lines are regressions for pursuit with body stationary. Dotted lines are regressions for pursuit during body movement. 56 of time lag has served to support the validity of the pur- suit model, and the conclusion of decreased time lag during body movement. 3. Are the data normally'distributed around their meansF Use of parametric statistics presupposes that the data are normally distributed around their means. A Kolmogorov- Smirnov test (Sokal and Rohlf, 1969) was used on several groups of data to test the null hypothesis that the data are normally distributed. The use of a t-test to conclude that the/average dif- ference between model gain and measured gain is not signifi- cantly different from d = 0.00, assumes that the underlying population of gain differences is normally distributed. For pure pursuit data, with observations more than five standard deviations from the mean trimmed, dmax L3370) was less than the critical value (.3376 for n = 18, °5 = .05) leading to a failure to reject the hypothesis of normally'distributedi data. For combined stimulation, however, the data distribution was found to be significantly different from normal (d a .3505, critical value =1.3376, n = 15, “=- max .05). For this population of gain differences, the variance is probably larger, than it would be if the data were normally distributed. This makes it more difficult to re- ject 30:3 = 0.00. Performances of individual chameleons might be expected to be normally distributed around their own means; and if a large enough group of chameleons were tested, the data 57 should be normally distributed around a group mean. Susanks pursuit data are not significantly different from being nor- mally distributed (d = .3449, critical value = .45427, n max = 8,‘x=-.05), and her average gain difference is not sig- nificantly different from d = 0.00 (n = 8, d = -0.04, s5 = 0.02, t a 1.66, 05-».05). It is probable that not enough chameleons were tested, to make the group data normally distributed.in the case of combined stimulation (3 chame- leons for combined stimulation, compared to 5 for pursuit data). Other sampling problems may have contributed to dataks not being normally distributed: 1) the chameleons were not uniformly sampled, i.e., more data were taken from some than others: 2) the fit of the model might have a frequency dependence, and frequencies were not uniformly sampled (see Tables 1 and 2). In spite of these sampling problems, measured gain is significantly correlated with model gain (as reported in Chapters 2 and 3), and the ratio of measured gain/model gain (although not normally distributed) is not significantly different from 1.00 (n a 44, 37 =- 0.96, 33; = 0.04, t = 1.00, 05= .05). The use of a t-test to conclude that time lags were significantly shorter during body movement (Table 3-1) as- sumes that time lag differences are normally'distributed. Use of the Kolmogorov-Smirnov test on this population led to a failure to reject the hypothesis that the data are 58 normally distributed (d 0.2097, critical value = max .32733, n a 16, 0< =- .05). Although there are inherent problems in the use of heterogeneous data, and small sample sizes, it seems prob- able that the populations underlying the collected data are normally distributed and therefore, it is reasonable to use parametric statistics in this thesis. APPENDIX B LISSAJOUS ANALYSIS Lissajous analysis (Malmstadt and Enke, 1963) was used to calculate the gain and phase of the sinusoidal head movement that followed sinusoidal bait movement.(see Figures 6 and 9). This analysis makes use of the fact that the mathematical formula for an ellipse can be derived from two sine waves, one in the x dimension and another in the y dimension. For a bait movement with amplitude B and a head movement with amplitude H, the formulas for x and y coordi- nates are: x = 8 sin «it y =- H sin (wt '1'?) where A): 2TTf rad./sec., and ‘¥= Z’Nf At rad. f is the frequency of the sinuosoidal bait and head movements ( in cycles/sec.), and At "is the time lag of the head movement (in seconds). Figure 16 shows bait and head positions cross plotted for points in time during a typical pursuit head movement. The data were fitted (by eye) with an ellipse. Gain was determined, as shown, by taking the ratio of maximum y and x positions, H/B. Phase lag,9’, was determined by using the x and y equations at the point where sin A) t = 0, and solving for 1’: 59 60 Gain = H)/B Phase = sin" Yo/ H Figure 16: Gain and Phase Calculation. Head (H) and bait (B) positions are cross plotted to determine gain and phase by the Lissajous method. 61 x0 = 0 Y0 = H sin‘V and thus 3’: sin"1 yo/H. For a given set of data, repeated use of this graphi- cal, Lissajous method showed that the error of the technique (Taylor, 1982) was 1,0.02 for gain and i 0.05 rad. for phase. BI BLIOGRAPHY BIBLIOGRAPHY .Allen, Mary E. Department of Animal Science, Michigan State University. Barr, C. C., L. W. Schultheis, and D. A. Robinson. Voluntary, non-visual control of the human vestibulo- ocular reflex. Acta Otolaryngol. 81: 365-375, 1976. Beer, F. P. and E. R. Johnstion, Jr. Vector Mechanics for Engineers. (U.S.A.: McGraw-Hill, Inc.), 1977. Bellairs, A. The Life of Reptiles. (New York: Universe Books),1970. Bernstein, N. A. The coordination and regulation of move- ments. (London: Pergamon Press), 1967. Bizzi, E., R. E. Kalil, and V. Tagliasco. Eye-head coordi- nation in monkeys: evidence for centrally patterned organization. Science 113: 452-454, 1971. Bizzi, E., A. Polit, and P. Morasso. Mechanisms underlying achievement of final head position. J. Neurophysiol. 12 (2): 435-444, 1976. Bizzi, E., P. Dev, P. Morasso, and A. Polit. Effect of load disturbances during centrally initiated movements. J. Neurophysiol. 41 (3): 542-556, 1978. Bock, O. Non-linear interaction of the vestibular and eye tracking system in man. Exp. Brain Res. 41: 461-464, 1982. Gans, C. The Chameleon. Nat. Hist. N.Y. 15, (4): 52-59, 1967. Goodman, D. And J. A. S. Kelso. Exploring the functional significance of physiological tremor: a biospetro- scopic approach. Exp. Brain Res. 42: 419-431, 1983. Greene, P. H. Why is it easy to control your arms? J. Motor Behavior 14 (4): 260—286, 1982. Harkness, L. Chameleons use accommodation cues to judge distance. Nature 261: 346-349, 1977. 62 63 Harkness, L., and H. C. Bennet-Clark. The deep fovea as a focus indicator. Nature 212: 814-816, 1978. Hann, V., B. Cohen, and L. R. Young. Visual-vestibular interaction in motion perception and the generation of nystagmus. Neurosciences Res. Prog. Bull. 18 (4), 1980. Lanchester, B. S., and R. F. Mark. Pursuit and prediction in the tracking of moving food by a teleost fish (Winters: sbilmlanurus). J. Exp. Biol. 6.1: 627-645, 1975. Lara, R., M. A. Arbib, and A. S. Cromarty. The role of the tectal column in facilitation of amphibian prey- catching behavior: a neural model. J. Neurosci. 2 (4): 521-530, 1982. Lisberger, S. G., C. Evinger, G. W. Johanson, and A. F. Fuchs. Relationship between eye acceleration and retinal image velocity during foveal smooth pursuit in man and monkey. J. Neurophysiol. 45, (2): 229-249, 1981. Malmstadt, EL.‘V. and C. G. Enke. Electronics for Scientists: Principles and Experiments for Those Who Use Instruments. W. A. Benjamin, Inc., New York, 1963. Mates, J. W. B. Eye movements of African chameleons: spontaneous saccade timing. Science 122: 1087-1089, 1978. Meyer, D. L., W. Graf, and U. V. Seydlitz-Kurzbach. The role of integrators in maintaining actively assumed abnormal posutres. J. Comp. Physiol. 111: 235-246, 1979. Mivart, S. G. On the myology of Chameleon mil. Proc. 2001. Soc. Land. 850-890, 1870. Murphy, R. L. The most amazing tongue in nature. Nat. Pratt, D.VL. Saccadic eye movements are coordinated with head movements in walking chickens. J. Exp. Biol. 21: 217-223, 1982. Robinson, D. A. Eye movement control in vertebrates. In: Function and Formation of Neural Systems. ed. G. S. Stent (Berlin: Dahiem Konferenzen), 179-195, 1977. Snedecor, G. W. and W. G. Cochran. Statistical Methods. Sixth Edition. (Ames, Iowa: Iowa State University Press), 1971. 64 Sokal, R. R. and F. J. Rohlf. Biometry (San Francisco: W.H. Freeman and Co.), 1969. Steel, R. G. D. and J. H. Torrie. Principles and Procedures of Statistics: A Biometrical Approach. 2nd Edition. (New York: McGraw-Hill Book Co.), 1980. Taylor, J. R. An Introduction to Error Analysis: The Study of Uncertainties in Physical Measurements. (Mill Valley, Ca1.: University Science Books), 87-89, 1982. Walls, G. L. The Vertebrate Eye and its Adaptive Radiation. (Bloomfield Hills, Michigan: The Cranbrook Press), 1942. "I7'1?(11111711111111?