- Email: [email protected]
Dynamic adaptation of the blind pointing characteristic to stepwise lateral tilts of body, head, and trunk
99
Behavioural Brain Research, 30 (1988) 99-110 Elsevier BBR00825
Research Reports
Dynamic adaptation of the blind pointing characteristic to stepwise lateral tilts of body, head, and trunk Dietmar Ott and Rolf Eckmiller Division of Biocybernetics, Department of Biophysics, University of Diisseldorf (F.R.G.) (Received 8 August 1987) (Revised version received 28 January 1988) (Accepted 2 February 1988)
Key words: Blind pointing; Sensorimotor; Coordinate transformation; Otolith; Neck; Man
Blind pointing (i.e. pointing to visual targets without seeing the pointing arm) was investigated in normal subjects in response to stepwise lateral tilts (20 °) of the body, head, and trunk. Blind pointing positions of the right index finger on the outer surface of a hemispherical screen were measured relative to the positions of visual targets that were presented along a horizontal line ( + 30 ° in head coordinates) on the inner screen surface, thus yielding a blind pointing characteristic (BPC). (1) BPC is highly reproducible and can be subdivided into separate branches for the ipsi- and contralateral hemifields. These branches are rotated relative to tile target line by individually different BPC angles ni and %. (2) Iti exhibits characteristic time courses (measured within 10 min following a stepwise tilt) for each paradigm. (3) Body tilt (left ear down) causes a steplike increase in ni of up to 14 ° ; body tilt (right ear down) causes a step-like decrease in ni to about zero. (4) Trunk tilt (right shoulder down) produces a gradual decrease in ~i of up to 6 ° (average time constant Z,r = 5 min); trunk tilt (left shoulder down) produces a gradual increase in ni of up to 4 °. (5) Head tilt (left ear down) causes an increase in ni of up to 9 ° followed by a gradual decrease (average time constant TH = 6 min); head tilt (right ear down) causes a step-like decrease in iti with unsignificant further changes. These findings are discussed in terms of a neural sensorimotor coordinate transformation process receiving separate, dynamic otolith and neck afferent influences. INTRODUCTION
Pointing or reaching within the grasping space (with or without visual guidance) is based on internally generated motor programs probably requiring several neural representations of space coordinates at sensory and motor levels. Recent behavioral studies in trained monkeys 1,37 and human subjects in an upright position 2,31-33 have addressed several aspects of the coordinate transformation (CT) of target positions in visual space into pointing positions in grasping space without visual feedback from t h e pointing arm (blind pointing).
The position of the arm in space depends on trunk position and is related to retinal coordinates via eye position relative to head and head position relative to trunk. Thus, we assume that a given blind pointing movement requires neural space representations on various levels and intermediate coordinate transformations. Fig. 1 schematically illustrates the various possible steps involved in this multi-coordinate transformation process. Possible distortions of the various neural maps in this scheme are indicated by the curved circumference. Blind pointing at a visual target position P in physical space is indicated by the open circle. When the target moves
Correspondence: D. Ott, Division of Biocybernetics, Department of Biophysics, University of Dilsseldorf, F.R.G. 0166-4328/88/$03.50 © 1988 Elsevier Science Publishers B.V. (Biomedical Division)
100 from P to X, the new blind pointing positioi; is given by an open cross. For this purpose, the subject has to first fixate the new target. This requires that retinal target eccentricity has to be transformed (CT]) into an appropriate oculomotor command. Accordingly, P is projected onto P~ and X onto X] in a postulated neural 'eye position map' (Fig. 1). Any point within this map corresponds to a given eye position relative to head position. Horizontal (He) and vertical (Ve) coordinates of P~ and X~ are assumed to be represented by horizontal and vertical oculomotor efference copy signals. If a re-fixating eye movement from P to X occurs during head tilt in space (we confine these considerations to roll tilt only), the eye position map rotates by an angle b defined as the difference between the head tilt angle and the tilt-induced ocular counter-roll angle 7. CT 2 in Fig. 1 transforms space coordinates from the eye position map onto a postulated
neural 'internal space map' with P2 and X 2. Angle ~ gives the rotation angle of this map between the gravitation vector and the 'subjective vertical '23. We assume that C T 2 incorporates tilt information from various sensory modalities, e.g. vestibular-otolith signals (indicating head tilt) ]4, somatosensory signals (indicating trunk tilt and head tilt) ~5, as well as tilt information derived from visual features of a structured panoramic background ('visual capture') 24. Thus, the internal space map is a head-centered neural representation of physical space similar to the 'common spatial reconstructor' ]6. The internal space map (Fig. 1) is subsequently projected onto a postulated neural 'pointing motor map' by means of C T 3. Any point within this map represents a desired pointing position. A c c o r d i n g l y , C T 3 is assumed to incorporate the tilt angle of the trunk relative to the head, information probably provided by neck-proprioception.
physical space Y
X
eye position map
/~~
~ ' ~
internal space map
MULTIMODAL TILT
pointingmotor map
/ NECK TILT
Fig. 1. Suggested scheme representing the possible neural space representations and coordinate transformations underlying sensorimotor integration in blind pointing.
101 Angle ~/ indicates the rotation of the pointing motor map relative to the gravitation vector and is def'med as the difference between the tilt angle of the trunk relative to the head and angle e. F i n a l l y , C T 4 (Fig. 1) describes the transformation of desired pointing positions (e.g. P3 and X3) into actual blind pointing positions in physical space (e.g. open circle and open cross). It was the goal of the present study to investigate the dynamic changes in blind pointing performance of normal human subjects in response to specific stepwise tilt paradigms: (A) body relative to the gravitation vector, (B) head relative to the upright trunk, and (C)trunk relative to the upright head (Fig. 2). We will provide evidence for separate neckproprioceptive and vestibular-otolith influences with different dynamic properties on the sensorimotor CT of visual target into blind pointing positions. A short summary of this paper has been presented elsewhere27.
BO DY
MATERIALS AND METHODS
Subjects were seated comfortably in a swing chair with the head held in a headrest and the trunk fixated by softly padded supports. Upper (fixating the head) and lower (fixating the trunk) chair parts could be rotated independently about the horizontal X-axis passing approximately through the head joint, thus allowing static lateral tilt (tilt angle 20 °) of either the head relative to the upright trunk, the trunk relative to the upright head, or the whole body relative to the gravitation vector. Visual targets were presented sequentially along a horizontal line (relative to head coordinates) as one of 21 red light-emitting diodes (LEDs) of 1 mm diameter on the inner surface of a hemispherical screen (radius: 23 cm, thickness: 2.5 cm). The screen was attached to the upper chair part in front of the subject's eyes and was centered on the nasal root. Thus, tilt of the upper chair part rotated both the subject's head and the
HE AD
/
TRUNK
/
I ° - - ~ I
g
g
~-~~
g
Fig. 2. Schematic representation of the static tilt paradigms in which blind pointing performance was measured: body tilt (left ear down) relative to the gravitation vector, head tilt (left ear down) relative to the upright trunk, trunk tilt (right shoulder down) relative to the upright head, and (not shown) trunk tilt (left shoulder down) relative to the upright head. Typical blind pointing characteristics (BPC: horizontal and vertical coordinates of blind pointing positions (open symbols) relative to visual target positions (filled symbols)) of one subject are shown as insets, ni indicates the rotation angle of the ipsilateral BPC branch relative to the horizontal line (in head coordinates) of the visual targets.
102 screen such that the line of the visual targets was always horizontal relative to head coordinates. At each visual target position (0, + 10, +20 and +30 ° of visual angle), LEDs were presented 3 times in a computer-controlled (Apple II) order (a sequence of 7positions being repeated 3 times). The inside of the screen was dimly illuminated by about 1 cd/m 2. The subjects wore ski-goggles with milky glass such that only the lit LED could be seen in an otherwise unstructured surround yielding no visual reference frame and preventing the sight of the pointing arm (blind pointing). Horizontal and vertical coordinates of the final pointing positions of the pointing index finger tip were measured on the outer surface of the screen by means of a modified joy stick device. Specifically, subjects wore a thimble over their pointing index finger. This thimble was attached to the end of a maneuverable stick which was in turn attached to a cardan joint. Calibrated precision potentiometers recorded the horizontal and vertical angles of the movement of the stick. These angles were on-line converted into spherical blind pointing coordinates (centered on the subject's nasal root) by means of a computer algorithm. (Consequently, BPC is a cartesian projection of spherical coordinates.) Resolution of finger position measurement was limited only by the digitization (8 bit A/D), yielding 0.3 ° for both horizontal and vertical direction. Subjects were instructed to point as accurately as possible to the binocularly presented LEDs. They indicated the termination of a blind pointing movement by pressing a button with the nonpointing hand. This signal triggered the acquisition of the blind pointing position, termination of the LED, and illumination of the next LED with a time delay of 500 ms. Subjects performed blind pointing without lowering the pointing arm between target presentation. No time constraints were imposed on this procedure, allowing the subjects to operate at their individual pace. A single measurement cycle (with 21 target positions) typically took 1 min without causing any fatigue to the pointing arm. The blind pointing characteristic (BPC), i.e. horizontal and vertical coordinates of blind
pointing positions relative to the visual target positions, developed on a monitor screen in the course of the measurement cycle and was stored on a magnetic disc for later analysis. Thirty-eight normal healthy subjects took part in a short version of the experiment. They performed one measurement cycle (21 target presentations, right hand pointing) in an upright position, were then tilted (body tilt, left ear down) and performed a second measurement cycle in the tilted position after an adaptation period of 1 min. Four of these subjects (25-43 years of age) participated in a longer version of the experiment. For each experimental session, subjects performed a sequence of 6 successive measurement cycles, each separated from the onset of the subsequent one by 3-min breaks. The first two measurement cycles were always made in the upright position. The last 4 were made following a stepwise tilt of the body (left ear down), head (left ear down), trunk (right shoulder down), or trunk (left shoulder down), (see Fig. 2). The first measurement cycle in tilted position was again initiated 1 min post-tilt. For comparison, right hand pointing of several subjects was quantitatively evaluated during tilt in the opposite direction (right ear down during body or head tilt). Also, pointing performance with the left pointing hand was studied in several subjects in order to examine the possible influence of 'handedness' on the results. RESULTS
(A) Reproducibility Three representative BPCs (obtained 1 min after adaptation to a stepwise tilt during the various paradigms) for one of the subjects undergoing the long version of the experiments (right hand pointing) are presented in Fig. 2. Open symbols indicate blind pointing positions corresponding to the various visual target positions (filled symbols). As can be seen, the reproducibility of repeated blind pointing movements (indicated by the closely clustered triplets) is very good. Standard deviations calculated for each triplet of blind pointing positions were independent of the visual
103 hemifield of target presentation and target eccentricity ( + 30°). Standard deviations were averaged across all target positions, thus yielding the range of blind pointing variability. The ranges of all obtained BPCs (all subjects and paradigms) varied between 0.8 and 3.9 ° . The range values of measurement cycles in upright position (0.8-3.4 °) were not significantly ( P > 0 . 1 ) different from those during tilt (1.0-3.9°). The range of repeated measurement cycles (within a given sequence) was highly reproducible and varied only fiainimally when compared at different days of trial. Blind pointing movements in response to the appearance of a new LED were performed with approximately straight trajectories and without large correctional adjustments upon attainment of the subjectively localized target position.
(B) Offset As can be depicted qualitatively from the BPCs within Fig. 2, the entire string of blind pointing positions was found to be individually shifted in horizontal as well as in vertical direction relative to the line of the visual targets. These offsets showed large interindividual differences. For measurements in upright position, offsets (calculated as shifts of the primary blind pointing triplet relative to the primary target position) were found in the range between 3.6 ° up and 7.5 ° down and between 13.1 ° left and 8.5 ° right (right hand pointing). A more detailed analysis of the static offsets will be the subject of a separate study.
(C) BPC dichotomy The most important finding was a significant separation of the BPC (usually at the blind pointing triplet corresponding to the primary LED position, see Fig. 2) into two BPC 'branches' for the ipsi- and contralateral pointing hemifields (ipsi- and contralateral relative to the pointing finger). The relationship between the centers of blind pointing triplets within either hemifield (positions 0 ° to 30 ° right, and 30 o left to 0 °) were approximated by linear regression lines. The slopes of these lines def'med the BPC angle rri (for the ipsilateral pointing hemifield) and zr¢ (for the contralateral pointing hemifield) by
which the ipsi- and contralateral BPC branches were rotated with respect to the line of targets. Positive (negative) angles n indicate that the vertical error (relative to the line of targets) of blind pointing positions continuously increased (decreased) in the ipsi-/contralateral hemifield with increasing pointing eccentricity. In the upright position, rri ranged from - 3 to 19 ° with an average value of ~i = 3 o (n = 38 subjects). In tilted position (body tilt left ear down), rq ranged from 0 to 34 o with an average value of ni = 6.6°. The average values of rcc were about zero. Fig. 3 shows the distribution of the differences (Arri) between ni during static body tilt (left ear down, right hand pointing) and in the upright position for the 38 subjects. Bin width is 1 °. Numbered squares represent the average values of the 4 main subjects. In 35 subjects, zri clearly increased by an average amount of 3.6 ° following the tilt, thus confirming our previous findings 8.
(19) Time courses of BPC angles zr Fig. 4 presents the time courses of n for subject 1 (see Fig. 3) in the upright position (left values within each diagram) and following a stepwise tilt (right values) for each of the 4 tilt paradigrns. The upper set of 6 curves within each diagram represents the time courses of rq for the 6 repeated trials (indicated by different symbols) to indicate typical intra-individual variability. The heavy line connects the average values of zri. For N
-2
2
4
6
N
10
12
A 7~; (deg) Fig. 3. Histogramshowingthe distributionofthe tilt-induced changesin BPC anglen~for 38 subjects(righthand pointing). The abcissa shows the differences(A~) between n~ during static bodytilt (leftear down)and in the uprightposition. Bin width is 1°. Numberedsquares representthe averagevalues for the 4 main subjects. Positivevalues indicate an increase in Iti during tilted as compared to upright position.
104
A
B
i
J
30"
30"
20-
20"
o v
o---_____o
10
10-
._.-O
O'-
-4
-2
I
I
I
I
2
4
6
8
I
I
-4
-2
! a
10
2
4
6
8
10
t (min)
-10
-1C
D
C 301
10
30
¸
"
10'
~>~
,/
I
I
I
I
I
I
-4
-2
2
4
6
8
I ----
10
/
I
-4
I
I
I
I
-2
2
4
6
• ....... • ......
-1C
I
s- ~, ~
8;" 10 Os"
-10
Fig. 4. Time courses of BPC angles nofsubject 1 in the upright position(left values within each diagram)and following astepwise tilt (right values): A: body tilt (left ear down). B: head tilt (left ear down). C: trunk tilt (right shoulder down). D: trunk tilt (left shoulder down). The set of curves represents the time courses of ni for 6 repeated trials (indicated by different symbols). The heavy line connects the average values of hi. For nc (broken line), only the average time courses are displayed.
105 nc (broken line), only the average time course is displayed. The average values of rri and rr~ for the trials in the upright which preceded the tilt were similar, yielding about 12 ° and 4 ° respectively. In response to a stepwise tilt, ~, and to some degree zr¢, were found to change with characteristic time courses depending on the paradigm. Stepwise body tilt (left ear down, right hand, Fig. 4A) caused a steplike increase in rq of up to about 14 ° immediately after tilt, with a subsequent gradual decrease during the following 9 min. No significant changes occurred for rcc. Stepwise head tilt (left ear down, right hand, Fig. 4B) led to a transient increase in rq of up to 6 °, followed by a gradual decrease, with an average time constant zri of 5 min, to about the pre-tilt value, fro decreased only slightly following tilt. Stepwise trunk tilt (right shoulder down, right hand, Fig. 4C) gradually decreased rri from the
A
t
~
.........r ...~
pre-tilt value up to about 6 ° with an average time constant Z,r of 7 min. n c also decreased following a similar time course. Conversely, stepwise trunk tilt (left shoulder down, right hand, Fig. 4D) gradually increased both rri (Tmal' increase about 1.5 °) and rrc. Previous studies 26 on more than 10 subjects had shown that right hand pointing during body tilt (right ear down) caused a step-like decrease in rri immediately following tilt. ni typically yielded positive values close to zero. In conjuction with the present study, the time course of zri following right ear down tilt was monitored for two subjects (numbers 2 and 3 in Fig. 5). It was found that body tilt (right ear down) yielded iti of 1-2 o which remained constant for the measurement episode of 1-10 min following tilt. For comparison, right hand pointing was also tested following head tilt (right ear down). Again ni dropped to smaller values (although not as far
B
:
•
.2
~
.........
~°,
........
10 /
..... ~ ~ _ - ~ 4 -4
_ -2
I
2
4
6
8
10
C
D
.... ~
• ---------'~
o . . . . ~ / ~ ~ - ~ .......... I ""
-2 ;
7 .........
. . . . .
:
"
' ........:t
:
'
'--Z'4--
Fig. 5. Average time courses of BPC angle 7ri for 4 representative normal subjects (1-4) and tilt paradigms A - D (see Fig. 4).
106 for body tilt) and remained essentially constant throughout the measurement episode of 10 min. The average time courses of n i for 4 representative normal subjects (1-4) for paradigms A - D are summarized in Fig. 5. As can be seen, the qualitative shape of the time courses was similar in all subjects, whereas the pre-tilt values and the magnitude of the tilt-induced changes of 7q showed considerable inter-individual differences. The average time constant rH was 6min (S.D. = 4.2 min); the average time constant cawas 5 min (S.D. = 3.4 min). Measurement sequences in the upright position (without tilt) yielded flat time courses with nonsignificant variations. Occasionally subjects were restored to the upright position at the end of a measurement sequence and additional measurement cycles were performed to see whether the tilt-induced changes in n i were reversible. When returned from head or body tilt, ni typically decreased during the first measurement cycle in the upright position and subsequently reached the initial pre-tilt value within the next one or two measurement cycles. When returned from trunk tilts, the return of n i was more gradual, reaching pre-tilt values only after several minutes.
(E) Comparison of left versus right hand pointing It has been reported in a preliminary study 26 that body tilt (right ear down) combined with left hand pointing yielded just mirror-symmetrical results as compared to the right hand/body tilt (left ear down) constellation. Therefore, as a control, we tested pointing performance with the left pointing hand and paradigms A - D for several subjects. It was found that the time courses of n were indeed mirror-symmetrical to those described in part (D) for all paradigms A - D . The individual magnitude of the tilt-induced changes in ni as well as the time constants ZH and Vr corresponded very closely to the values found for the right pointing hand. (F) Postural and visual perceptions All subjects reported postural adaptation during prolonged tilt. When returned to the upright position following sustained body tilt, subjects felt
as though they were being tilted in the opposite direction. Similarly, in the head and trunk paradigm our subjects reported tilt adaptation which diminished gradually with individual time constants ranging from less than one up to several minutes. Although only a single LED was visible at a time, all subjects were able to make approximate estimates of the subjective rotation of the line of visual targets relative to their head. Similar to BPC dichotomy, most subjects also reported a dichotomy in the visual perception of the targets (the contralateral targets were perceived along a line which was almost horizontal relative to the head, whereas the ipsilateral targets were perceived to appear along a line which was rotated relative to the head horizontal). Tilt-induced perceptual rotations of the ipsi- or contralateral half of the target line were mostly coupled with synergistic changes in 7zi and no. DISCUSSION
The present study is based on the novel finding8 that the BPC can typically be subdivided into separate, approximately straight 'branches' for the ipsilateral and the contralateral pointing hemifield. More specifically, the contralateral branch was tyically nearly parallel with the line of visual targets (always horizontal in head coordinates), whereas the ipsilateral BPC branch was rotated by an angle hi, indicating a continuous increase in the vertical error (relative to target line) of blind pointing positions with target eccentricity into the ipsilateral hemifield. This highly reproducible dichotomy of the BPC into two branches is in general agreement with neurophysiological5' ~8.19, behavioural TM, and clinical studies 28, suggesting a hemispherical subdivision of sensori-motor processing in primates. With regard to our suggested scheme for the sequence of coordinate transformations (Fig. 1), it is worth emphasizing that subjects also reported a dichotomy in the perception of the visual targets.
Otolith and neck afferent influence on BPC Stepwise body tilt (left ear down, right hand pointing, Figs. 4 and 5) caused an immediate,
107 stepwise increase in 7q of up to 14 ° with only gradual (if any) subsequent changes. In contrast, body tilt (right ear down) caused a step-like decrease in ni to about zero. As tilt-induced vestibular canal activity should have reached its new steady state following a post-tilt adaptation period of 1 min 36, and no visual spatial reference cues were available, this step-like increase of rri is assumed to be caused by otolith function. This hypothesis rests on neurophysiological data regarding the directional selectivity and static impulse rate modulation of otolith afferents upon adequate stimulation around the roll axis 9'3°. In these terms, the gradual, post-tilt changes in 7~i could reflect long-term otolith adaptation during sustained stimulation 9"12"35. Further evidence supporting our hypothesis of otolith-induced changes is provided by the close correspondence between changes in 7~ and the ocular countertorsion reflex 7. Studies with unilateral labyrinthectomized monkeys 17 and patients with selective otolith dysfunction 4,25 indicate that following lateral tilt, the lower otolith plays a dominant role in the corresponding ocular countertorsion. Preliminary blind pointing studies with unilateral labyrinthine-deficient patients in our own laboratory 13 showed that ni and 7L were typically outside the normal range for lateral tilt towards the deficient ear but normal for tilt in the opposite direction. Stepwise trunk tilt (right shoulder down, right hand pointing, Figs. 4 and 5) caused a gradual decrease in ni (up to 6 ° within 10 min) and to some degree also in no. Stepwise trunk tilt (left shoulder down) caused the reverse effect (gradual increase in 7ti of up to 4 °). As the head was held stationary in an upright position (no otolith stimulation) and the trunk was supported by soft supports (minimizing stimulation of skin receptors), these gradual changes in the BPC are assumed to be caused by neck afferent signals. This hypothesis is compatible with neurophysiological data yielding antagonistic impulse rate modulation of neck afferents during lateral tilt of the trunk relative to the stationary head 6,22,34. The individual time constants (zH and Zr) of the gradual decrease in ni were typically in the range of several minutes, in agreement with the time constants found for the
dynamic changes in the 'subjective vertical' in response to stepwise lateral head tilt24'39). However, the neural time constants of neck afferents are typically in the range of a few seconds 34. Stepwise head tilt (left ear down, right hand pointing, Figs. 4 and 5) led to a transient increase in 7q of up to 9 ° (less than during body tilt, left ear down) which then gradually decreased again, 'finally' arriving at or below the pre-tilt value, reo exhibited only a gradual decrease without a transient response. In contrast, head tilt (right ear down) produced a step-like decrease in 7ti (again less than during body tilt, right ear down) followed by only gradual further changes depending on the subject. Since head tilt involves both otolith stimulation (as during body tilt) and neck stimulation (as during trunk tilt), this PD-type time course would be expected assuming a linear interaction between otolith and neck signals. This notion is supported by additional experiments: combined head tilt (20 o left ear down) and trunk tilt (40 ° left shoulder down) yielded an initial step-like increase in ni (as expected from otolith stimulation) followed by a subsequent further increase (as expected from neck stimulation). Our present findings of separate, dynamic otolith and neck influences on the BPC are in agreement with the neural otolith-neck convergence pattern 3'11.29, the coordination of static limb reflex balance in postural control 2°'21, and the tilt-induced changes of the 'subjective visual vertical '38. Control experiments with the left pointing arm in the reverse tilt paradigms, i.e. body/head tilt (right ear down), trunk tilt (left/right shoulder down), yielded mirror-symmetrical results, thus supporting our hypothesis of a hemifield-specific influence of tilt signals on the neural control of blind pointing movements for each hand. The entire string of blind pointing positions was found to be individually offset relative to the line of the visual targets in vertical as well as in horizontal direction. These shifts of the BPC could be interpreted as offsets of the various postulated neural space maps, see Fig. 1. However, a detailed analysis of this f'mding will be the subject of separate study. On the basis of theoretical considerations, we proposed (see Introduction) that blind pointing
108 involves processing of space coordinates along a chain of various internal neural 'space maps', thus involving intermediate coordinate transformations (Fig. 1) which are assumed to incorporate multimodal tilt information, e.g. from vestibular, visual, proprioceptive receptors. Since blind pointing is performed without visual feedback from the pointing arm (and hand), tilt-induced changes in the BPC possibly reflect the influence of otolith versus neck afferent signals on the neural coordinate transformations, influences which remain obscured (or compensated) under normal closed-loop visual conditions. Our findings suggest that otolith influence on the BPC is asymmetrical, causing significant changes predominantly in the ipsilateral pointing hemifield (rti), whereas the contralateral hemifield (rtc) is rather insensitive to lateral tilt. In contrast, neck afferent stimulation was found to symmetrically influence the pointing hemifields, although ipsilateral changes were typically more prominent than contralateral ones. Both the BPC and the subjective perception of the visual targets underwent marked adaptation upon stepwise change in body position. This is taken as support for the hypothetized existence of an 'internal space map' (Fig. 1) which is an internal representation of space coordinates accessible to both motor program generation as well as to visual perception ~6. Our closed-loop scheme (Fig. 1) of the various coordinate transformations might imply a serial processing of space coordinates through various different neural space maps. We want to emphasize that this sequential arrangement has been chosen for clarity to help the reader to separate and identify the various logical steps the CNS probably has to cope with in the complex task of sensorimotor integration underlying blind pointing. Blind pointing does not necessarily require a serial coordinate transformation process. Rather, it is plausible to assume the existence of a single neural network structure incorporating a continuously updated (via multimodal sensory input) neural space representation, thus making sensorimotor coordinate transformation a parallel process. Our hypothesis of otolith and neck afferent influence on the BPC has been critically assessed in the following ways.
(1) During body tilt, blind pointing movements were the same as those required in upright position but had to cope with an increasing gravitational pull in the direction of the pointing arm. Accordingly, ni might be the result of an overcompensation in motor effort in an attempt to overcome these constraints. We have tested this possibility by supporting the pointing arm with an elastic band from above and found no significant changes in the blind pointing performance. (2) Control measurements with randomized visual target positions insured that our findings were only minimally affected by predictability. (3) The dynamic changes in the BPC following body and trunk tilt might be caused by stimulation of cutaneous receptors in the trunk rather than by otolith and neck signals. However, the following arguments can be made against this possibility. (i) The individual time constants ~. and Vr correspond very closely for a given subject, indicating that the same (neck-) receptor dynamics are involved. This close correspondence also reflects the high individual reproducibility which could hardly be explained by tilt information arising from skin receptors being stimulated in a non-specific manner. (ii) Blind pointing measurements during body tilt (right ear down, right hand pointing) caused an immediate step-like decrease in ni rather than a gradual decrease, although stimulation of skin receptors was similar to that during trunk tilt (right shoulder down). The sensorimotor coordinate transformation paradigm described here requires highly skilled performance and interaction of various sensory, central-processing, and motor activities. Due to the high reproducibility of the BPC and its reversible dependence on the various tilt paradigms, the blind pointing paradigm could therefore be developed into a sensitive indicator for sensorimotor coordinate transformation performance in human patients and normal subjects.
ACKNOWLEDGEMENTS
We thank Ms. A. Offermans for technical assistance. Thanks are also due to Dr. O. Bock for his experimental contributions and discus-
109 sions. Supported by the Deutsche Forschungsgemeinschaft, SFB 200/B10. REFERENCES 1 Bizzi, E., Accornero, N., Chapple, W. and Hogan, N., Arm trajectory formation in monkeys, Exp. Brain Res., 46 (1982) 139-143. 2 Book, O. and Eckmiller, R., Goal-directed arm movements in absence of visual guidance: evidence for amplitude rather than position control, Exp. Brain Res., 62 (1986) 451-458. 3 Boyle, R. and Pompeiano, O., Convergence and interaction of neck and macular vestibular inputs on vestibulospinal neurons, J. Neurophysiol., 45 (1981) 852-868. 4 Brandt, T. and Dieterich, M., Pathological eye-head coordination in roll: tonic ocular tilt reaction in mesencephalic and medullary lesions, Brain, 110 (1987) 649-666. 5 Brinkman, J. and Kuypers, H.G.J.M., Cerebral control of contralateral and ipsilateral arm, hand, and finger movements in the split-brain rhesus monkey, Brain, 96 (1973) 653-674. 6 Denoth, F., Magherini, P.C., Pompeiano, O. and Stanojevic, M., Responses of Purkinje cells of the cerebellar vermis to neck and macular vestibular inputs, Pfliigers Arch., 381 (1979) 87-98. 7 Diamond, S.G. and Markham, C.H., Ocular counterrolling-as an indicator of vestibular otolith function, Neurology, 33 (1983) 1460-1469. 8 Eckmiller, R., Ott, D. and Bock, O., Coordinate transformation of visual target position into blind pointing position depending on the input from otolith organs, 8th International Symposium on Posturography, Amsterdam, 1986, p. 31. 9 Fernandez, C. and Goldberg, J.M., Physiology of peripheral neurons innervating otolith organs of the squirrel monkey. I. Response to static tilts and to long-duration centrifugal force, J. Neurophysiol., 39 (1976) 970-984. 10 Fisk, J.D. and Goodale, M.A., The organization of eye and limb movements during unrestricted reaching to targets in contralateral and ipsilateral visual space, Exp. Brain Res., 60 (1985) 159-178. 11 Fredrickson, J.M., Schwarz, D.W.F. and Kornhuber, H.H., Convergence and interaction of vestibular and deep somatic afferents upon neurons in the vestibular nuclei of the cat, Acta Otolaryngol., 61 (1966) 168-188. 12 Gernandt, B.E., Otolithic influences on extraocular and intraocular muscles, In 5th Symposium on the Role of Vestibular Organs in Space Exploration, SP-314, Washington DC, NASA, 1973. 13 Greven, C., Eckmiller, R. and Ott, D., Unpublished observations, 1986. 14 Guedry, F.E., Orientation of the rotation axis relative to gravity: its influence on nystagmus and the sense of orientation, Acta OtolaryngoL, 60 (1965) 30-48. 15 Howard, J.P. and Templeton, W.B., Human Spatial Orientation, Wiley, London, 1966.
16 Jeannerod, M., How do we direct our actions in space? In A. Hein and M. Jeannerod (Eds.), Spatially Oriented Behaviour, Springer, Berlin, 1983. 17 Krejcova, H., Highstein, S. and Cohen, B., Labyrinthine and extralabyrinthine effects on ocular counter-rolling, Acta Otolaryngol., 72 (1971) 165-171. 18 Lawrence, D.G. and Kuypers, H.G.J.M., The functional organization of the motor system in the monkey. I: the effects of bilateral pyramidal lesions, Brain, 91 (1968) 1-14. 19 Lawrence, D.G. and Kuypers, H.G.J.M., The functional organization of the motor system in the monkey. II: the effects of lesions of the descending brainstem pathways, Brain, 91 (1968) 15-36. 20 Lindsay, K.W., Roberts, T.D.M. and Rosenberg, J.R., Asymmetric tonic labyrinth reflexes and their interaction with neck reflexes in the decerebrate cat, J. Physiol. (Lond.), 261 (1976) 583-601. 21 Magnus, R., KOrperstellung. In Monographien aus dem Gesamtgebiet der Physiologie der Pflanzen und Tiere, Springer Berlin, 1924. 22 Mergner, T., Anastasopoulos, D. and Becker, W., Neural responses to horizontal neck deflection in the group X region of the cat's medullary brainstem, Exp. Brain Res., 45 (1982) 196-206. 23 Mittelstaedt, M., A new solution to the problem of subjective vertical, Naturwissenschaften, 70 (1983) 272-281. 24 Mittelstaedt, M., The effect of visual on extraretinal information about the vertical: suppression or superposition? Proc. XXIII Int. Cong. Psychol., Acapulco, Mexico, Elsevier, Amsterdam, Netherlands, 1984. 25 Nelson, J.R. and Cope, D., The otoliths and the ocular counter-torsion reflex, Arch. Otolaryngol., 94 (1971) 40-50. 26 Ott, D., Eckmitler, R. and Book, O., Transformation of visual target coordinates into blind pointing coordinates depending on the angle of body roll, Neurosci. Lett. Suppl., 22 (1985) 149. 27 Ott, D., Eckmiller, R. and Bock, O., Otolith and neck influence on the blind pointing characteristic, Neurosci. Lett. Suppl., 26 (1986) 327. 28 Perenin, M.T. and Vighetto, A., Optic ataxia: a specific disorder in visuomotor coordination, In A. Hein and M. Jeannerod (Eds.), Spatially Oriented Behaviour, Springer, Berlin, 1983. 29 Pompeiano, O., Manzoni, D., Srivastava, U.C. and Stampacchia, G., Convergence and interaction of neck and macular vestibular inputs on reticulospinal neurons, Neuroscience, 12 (1984) 111-128. 30 Pompeiano, O., Macular input to neurons of the spinoreticulocerebellar pathway, Brain Res., 95 (1975) 351-368. 31 Prablanc, C., Echallier, J.F., Komilis, E. and Jeannerod, M., Optimal response of eye and hand motor systems in pointing at a visual target. I. Spatio-temporal characteristics of eye and hand movements and their relationship when varying the amount of visual information, Biol. Cybern., 35 (1979) 113-124.
110 32 Prablanc, C., Echallier, J.F., Jeannerod, M. and Komilis, E., Optimal response of eye and hand motor systems in pointing at a visual target. II. Static and dynamic visual cues in the control of hand movement, Biol. Cybern., 35 (1979) 183-187. 33 Prablanc, C., Pelisson, D. and Goodale, M.A., Visual control of reaching movements without vision of the limb. I. Role of retinal feedback of target position in guiding the hand, Exp. Brain Res., 62 (1986) 293-302. 34 Rubin, A.M., Liedgren, S.R.C., Milne, A.C., Young, J.A. and Fredrickson, J.M., Vestibular and somatosensory interaction in the cat vestibular nuclei, Pfliigers Arch., 371 (1977) 155-160. 35 Schoen, L., Mikroableitungen einzelner zentraler Ves-
36
37
38
39
tibularisneurone von Knochenfischen bei Statolithenreizen, Z. Vgl. Physiol., 39 (1957) 399-417. Schrader, V., Koenig, E. and Dichgans, J., The effect of lateral head tilt on horizontal post-rotatory nystagmus l and II and the Purkinje effect, Acta Otolaryngol., 100 (1985) 98-105. Taub, E., Goldberg, I.A. and Taub, P., Deafferentation in monkeys: pointing at a target without visual feedback, Exp. Neurol., 46 (1975) 178-186. Wade, N.J., Visual orientation during and after lateral head, body, and trunk tilt, Percept. Psychophys., 3 (1968) 215-219. Wade, N.J., Effect of prolonged tilt on visual orientation, Q. J. Exp. Psychol., 22 (1970) 423-439.
Behavioural Brain Research, 30 (1988) 99-110 Elsevier BBR00825
Research Reports
Dynamic adaptation of the blind pointing characteristic to stepwise lateral tilts of body, head, and trunk Dietmar Ott and Rolf Eckmiller Division of Biocybernetics, Department of Biophysics, University of Diisseldorf (F.R.G.) (Received 8 August 1987) (Revised version received 28 January 1988) (Accepted 2 February 1988)
Key words: Blind pointing; Sensorimotor; Coordinate transformation; Otolith; Neck; Man
Blind pointing (i.e. pointing to visual targets without seeing the pointing arm) was investigated in normal subjects in response to stepwise lateral tilts (20 °) of the body, head, and trunk. Blind pointing positions of the right index finger on the outer surface of a hemispherical screen were measured relative to the positions of visual targets that were presented along a horizontal line ( + 30 ° in head coordinates) on the inner screen surface, thus yielding a blind pointing characteristic (BPC). (1) BPC is highly reproducible and can be subdivided into separate branches for the ipsi- and contralateral hemifields. These branches are rotated relative to tile target line by individually different BPC angles ni and %. (2) Iti exhibits characteristic time courses (measured within 10 min following a stepwise tilt) for each paradigm. (3) Body tilt (left ear down) causes a steplike increase in ni of up to 14 ° ; body tilt (right ear down) causes a step-like decrease in ni to about zero. (4) Trunk tilt (right shoulder down) produces a gradual decrease in ~i of up to 6 ° (average time constant Z,r = 5 min); trunk tilt (left shoulder down) produces a gradual increase in ni of up to 4 °. (5) Head tilt (left ear down) causes an increase in ni of up to 9 ° followed by a gradual decrease (average time constant TH = 6 min); head tilt (right ear down) causes a step-like decrease in iti with unsignificant further changes. These findings are discussed in terms of a neural sensorimotor coordinate transformation process receiving separate, dynamic otolith and neck afferent influences. INTRODUCTION
Pointing or reaching within the grasping space (with or without visual guidance) is based on internally generated motor programs probably requiring several neural representations of space coordinates at sensory and motor levels. Recent behavioral studies in trained monkeys 1,37 and human subjects in an upright position 2,31-33 have addressed several aspects of the coordinate transformation (CT) of target positions in visual space into pointing positions in grasping space without visual feedback from t h e pointing arm (blind pointing).
The position of the arm in space depends on trunk position and is related to retinal coordinates via eye position relative to head and head position relative to trunk. Thus, we assume that a given blind pointing movement requires neural space representations on various levels and intermediate coordinate transformations. Fig. 1 schematically illustrates the various possible steps involved in this multi-coordinate transformation process. Possible distortions of the various neural maps in this scheme are indicated by the curved circumference. Blind pointing at a visual target position P in physical space is indicated by the open circle. When the target moves
Correspondence: D. Ott, Division of Biocybernetics, Department of Biophysics, University of Dilsseldorf, F.R.G. 0166-4328/88/$03.50 © 1988 Elsevier Science Publishers B.V. (Biomedical Division)
100 from P to X, the new blind pointing positioi; is given by an open cross. For this purpose, the subject has to first fixate the new target. This requires that retinal target eccentricity has to be transformed (CT]) into an appropriate oculomotor command. Accordingly, P is projected onto P~ and X onto X] in a postulated neural 'eye position map' (Fig. 1). Any point within this map corresponds to a given eye position relative to head position. Horizontal (He) and vertical (Ve) coordinates of P~ and X~ are assumed to be represented by horizontal and vertical oculomotor efference copy signals. If a re-fixating eye movement from P to X occurs during head tilt in space (we confine these considerations to roll tilt only), the eye position map rotates by an angle b defined as the difference between the head tilt angle and the tilt-induced ocular counter-roll angle 7. CT 2 in Fig. 1 transforms space coordinates from the eye position map onto a postulated
neural 'internal space map' with P2 and X 2. Angle ~ gives the rotation angle of this map between the gravitation vector and the 'subjective vertical '23. We assume that C T 2 incorporates tilt information from various sensory modalities, e.g. vestibular-otolith signals (indicating head tilt) ]4, somatosensory signals (indicating trunk tilt and head tilt) ~5, as well as tilt information derived from visual features of a structured panoramic background ('visual capture') 24. Thus, the internal space map is a head-centered neural representation of physical space similar to the 'common spatial reconstructor' ]6. The internal space map (Fig. 1) is subsequently projected onto a postulated neural 'pointing motor map' by means of C T 3. Any point within this map represents a desired pointing position. A c c o r d i n g l y , C T 3 is assumed to incorporate the tilt angle of the trunk relative to the head, information probably provided by neck-proprioception.
physical space Y
X
eye position map
/~~
~ ' ~
internal space map
MULTIMODAL TILT
pointingmotor map
/ NECK TILT
Fig. 1. Suggested scheme representing the possible neural space representations and coordinate transformations underlying sensorimotor integration in blind pointing.
101 Angle ~/ indicates the rotation of the pointing motor map relative to the gravitation vector and is def'med as the difference between the tilt angle of the trunk relative to the head and angle e. F i n a l l y , C T 4 (Fig. 1) describes the transformation of desired pointing positions (e.g. P3 and X3) into actual blind pointing positions in physical space (e.g. open circle and open cross). It was the goal of the present study to investigate the dynamic changes in blind pointing performance of normal human subjects in response to specific stepwise tilt paradigms: (A) body relative to the gravitation vector, (B) head relative to the upright trunk, and (C)trunk relative to the upright head (Fig. 2). We will provide evidence for separate neckproprioceptive and vestibular-otolith influences with different dynamic properties on the sensorimotor CT of visual target into blind pointing positions. A short summary of this paper has been presented elsewhere27.
BO DY
MATERIALS AND METHODS
Subjects were seated comfortably in a swing chair with the head held in a headrest and the trunk fixated by softly padded supports. Upper (fixating the head) and lower (fixating the trunk) chair parts could be rotated independently about the horizontal X-axis passing approximately through the head joint, thus allowing static lateral tilt (tilt angle 20 °) of either the head relative to the upright trunk, the trunk relative to the upright head, or the whole body relative to the gravitation vector. Visual targets were presented sequentially along a horizontal line (relative to head coordinates) as one of 21 red light-emitting diodes (LEDs) of 1 mm diameter on the inner surface of a hemispherical screen (radius: 23 cm, thickness: 2.5 cm). The screen was attached to the upper chair part in front of the subject's eyes and was centered on the nasal root. Thus, tilt of the upper chair part rotated both the subject's head and the
HE AD
/
TRUNK
/
I ° - - ~ I
g
g
~-~~
g
Fig. 2. Schematic representation of the static tilt paradigms in which blind pointing performance was measured: body tilt (left ear down) relative to the gravitation vector, head tilt (left ear down) relative to the upright trunk, trunk tilt (right shoulder down) relative to the upright head, and (not shown) trunk tilt (left shoulder down) relative to the upright head. Typical blind pointing characteristics (BPC: horizontal and vertical coordinates of blind pointing positions (open symbols) relative to visual target positions (filled symbols)) of one subject are shown as insets, ni indicates the rotation angle of the ipsilateral BPC branch relative to the horizontal line (in head coordinates) of the visual targets.
102 screen such that the line of the visual targets was always horizontal relative to head coordinates. At each visual target position (0, + 10, +20 and +30 ° of visual angle), LEDs were presented 3 times in a computer-controlled (Apple II) order (a sequence of 7positions being repeated 3 times). The inside of the screen was dimly illuminated by about 1 cd/m 2. The subjects wore ski-goggles with milky glass such that only the lit LED could be seen in an otherwise unstructured surround yielding no visual reference frame and preventing the sight of the pointing arm (blind pointing). Horizontal and vertical coordinates of the final pointing positions of the pointing index finger tip were measured on the outer surface of the screen by means of a modified joy stick device. Specifically, subjects wore a thimble over their pointing index finger. This thimble was attached to the end of a maneuverable stick which was in turn attached to a cardan joint. Calibrated precision potentiometers recorded the horizontal and vertical angles of the movement of the stick. These angles were on-line converted into spherical blind pointing coordinates (centered on the subject's nasal root) by means of a computer algorithm. (Consequently, BPC is a cartesian projection of spherical coordinates.) Resolution of finger position measurement was limited only by the digitization (8 bit A/D), yielding 0.3 ° for both horizontal and vertical direction. Subjects were instructed to point as accurately as possible to the binocularly presented LEDs. They indicated the termination of a blind pointing movement by pressing a button with the nonpointing hand. This signal triggered the acquisition of the blind pointing position, termination of the LED, and illumination of the next LED with a time delay of 500 ms. Subjects performed blind pointing without lowering the pointing arm between target presentation. No time constraints were imposed on this procedure, allowing the subjects to operate at their individual pace. A single measurement cycle (with 21 target positions) typically took 1 min without causing any fatigue to the pointing arm. The blind pointing characteristic (BPC), i.e. horizontal and vertical coordinates of blind
pointing positions relative to the visual target positions, developed on a monitor screen in the course of the measurement cycle and was stored on a magnetic disc for later analysis. Thirty-eight normal healthy subjects took part in a short version of the experiment. They performed one measurement cycle (21 target presentations, right hand pointing) in an upright position, were then tilted (body tilt, left ear down) and performed a second measurement cycle in the tilted position after an adaptation period of 1 min. Four of these subjects (25-43 years of age) participated in a longer version of the experiment. For each experimental session, subjects performed a sequence of 6 successive measurement cycles, each separated from the onset of the subsequent one by 3-min breaks. The first two measurement cycles were always made in the upright position. The last 4 were made following a stepwise tilt of the body (left ear down), head (left ear down), trunk (right shoulder down), or trunk (left shoulder down), (see Fig. 2). The first measurement cycle in tilted position was again initiated 1 min post-tilt. For comparison, right hand pointing of several subjects was quantitatively evaluated during tilt in the opposite direction (right ear down during body or head tilt). Also, pointing performance with the left pointing hand was studied in several subjects in order to examine the possible influence of 'handedness' on the results. RESULTS
(A) Reproducibility Three representative BPCs (obtained 1 min after adaptation to a stepwise tilt during the various paradigms) for one of the subjects undergoing the long version of the experiments (right hand pointing) are presented in Fig. 2. Open symbols indicate blind pointing positions corresponding to the various visual target positions (filled symbols). As can be seen, the reproducibility of repeated blind pointing movements (indicated by the closely clustered triplets) is very good. Standard deviations calculated for each triplet of blind pointing positions were independent of the visual
103 hemifield of target presentation and target eccentricity ( + 30°). Standard deviations were averaged across all target positions, thus yielding the range of blind pointing variability. The ranges of all obtained BPCs (all subjects and paradigms) varied between 0.8 and 3.9 ° . The range values of measurement cycles in upright position (0.8-3.4 °) were not significantly ( P > 0 . 1 ) different from those during tilt (1.0-3.9°). The range of repeated measurement cycles (within a given sequence) was highly reproducible and varied only fiainimally when compared at different days of trial. Blind pointing movements in response to the appearance of a new LED were performed with approximately straight trajectories and without large correctional adjustments upon attainment of the subjectively localized target position.
(B) Offset As can be depicted qualitatively from the BPCs within Fig. 2, the entire string of blind pointing positions was found to be individually shifted in horizontal as well as in vertical direction relative to the line of the visual targets. These offsets showed large interindividual differences. For measurements in upright position, offsets (calculated as shifts of the primary blind pointing triplet relative to the primary target position) were found in the range between 3.6 ° up and 7.5 ° down and between 13.1 ° left and 8.5 ° right (right hand pointing). A more detailed analysis of the static offsets will be the subject of a separate study.
(C) BPC dichotomy The most important finding was a significant separation of the BPC (usually at the blind pointing triplet corresponding to the primary LED position, see Fig. 2) into two BPC 'branches' for the ipsi- and contralateral pointing hemifields (ipsi- and contralateral relative to the pointing finger). The relationship between the centers of blind pointing triplets within either hemifield (positions 0 ° to 30 ° right, and 30 o left to 0 °) were approximated by linear regression lines. The slopes of these lines def'med the BPC angle rri (for the ipsilateral pointing hemifield) and zr¢ (for the contralateral pointing hemifield) by
which the ipsi- and contralateral BPC branches were rotated with respect to the line of targets. Positive (negative) angles n indicate that the vertical error (relative to the line of targets) of blind pointing positions continuously increased (decreased) in the ipsi-/contralateral hemifield with increasing pointing eccentricity. In the upright position, rri ranged from - 3 to 19 ° with an average value of ~i = 3 o (n = 38 subjects). In tilted position (body tilt left ear down), rq ranged from 0 to 34 o with an average value of ni = 6.6°. The average values of rcc were about zero. Fig. 3 shows the distribution of the differences (Arri) between ni during static body tilt (left ear down, right hand pointing) and in the upright position for the 38 subjects. Bin width is 1 °. Numbered squares represent the average values of the 4 main subjects. In 35 subjects, zri clearly increased by an average amount of 3.6 ° following the tilt, thus confirming our previous findings 8.
(19) Time courses of BPC angles zr Fig. 4 presents the time courses of n for subject 1 (see Fig. 3) in the upright position (left values within each diagram) and following a stepwise tilt (right values) for each of the 4 tilt paradigrns. The upper set of 6 curves within each diagram represents the time courses of rq for the 6 repeated trials (indicated by different symbols) to indicate typical intra-individual variability. The heavy line connects the average values of zri. For N
-2
2
4
6
N
10
12
A 7~; (deg) Fig. 3. Histogramshowingthe distributionofthe tilt-induced changesin BPC anglen~for 38 subjects(righthand pointing). The abcissa shows the differences(A~) between n~ during static bodytilt (leftear down)and in the uprightposition. Bin width is 1°. Numberedsquares representthe averagevalues for the 4 main subjects. Positivevalues indicate an increase in Iti during tilted as compared to upright position.
104
A
B
i
J
30"
30"
20-
20"
o v
o---_____o
10
10-
._.-O
O'-
-4
-2
I
I
I
I
2
4
6
8
I
I
-4
-2
! a
10
2
4
6
8
10
t (min)
-10
-1C
D
C 301
10
30
¸
"
10'
~>~
,/
I
I
I
I
I
I
-4
-2
2
4
6
8
I ----
10
/
I
-4
I
I
I
I
-2
2
4
6
• ....... • ......
-1C
I
s- ~, ~
8;" 10 Os"
-10
Fig. 4. Time courses of BPC angles nofsubject 1 in the upright position(left values within each diagram)and following astepwise tilt (right values): A: body tilt (left ear down). B: head tilt (left ear down). C: trunk tilt (right shoulder down). D: trunk tilt (left shoulder down). The set of curves represents the time courses of ni for 6 repeated trials (indicated by different symbols). The heavy line connects the average values of hi. For nc (broken line), only the average time courses are displayed.
105 nc (broken line), only the average time course is displayed. The average values of rri and rr~ for the trials in the upright which preceded the tilt were similar, yielding about 12 ° and 4 ° respectively. In response to a stepwise tilt, ~, and to some degree zr¢, were found to change with characteristic time courses depending on the paradigm. Stepwise body tilt (left ear down, right hand, Fig. 4A) caused a steplike increase in rq of up to about 14 ° immediately after tilt, with a subsequent gradual decrease during the following 9 min. No significant changes occurred for rcc. Stepwise head tilt (left ear down, right hand, Fig. 4B) led to a transient increase in rq of up to 6 °, followed by a gradual decrease, with an average time constant zri of 5 min, to about the pre-tilt value, fro decreased only slightly following tilt. Stepwise trunk tilt (right shoulder down, right hand, Fig. 4C) gradually decreased rri from the
A
t
~
.........r ...~
pre-tilt value up to about 6 ° with an average time constant Z,r of 7 min. n c also decreased following a similar time course. Conversely, stepwise trunk tilt (left shoulder down, right hand, Fig. 4D) gradually increased both rri (Tmal' increase about 1.5 °) and rrc. Previous studies 26 on more than 10 subjects had shown that right hand pointing during body tilt (right ear down) caused a step-like decrease in rri immediately following tilt. ni typically yielded positive values close to zero. In conjuction with the present study, the time course of zri following right ear down tilt was monitored for two subjects (numbers 2 and 3 in Fig. 5). It was found that body tilt (right ear down) yielded iti of 1-2 o which remained constant for the measurement episode of 1-10 min following tilt. For comparison, right hand pointing was also tested following head tilt (right ear down). Again ni dropped to smaller values (although not as far
B
:
•
.2
~
.........
~°,
........
10 /
..... ~ ~ _ - ~ 4 -4
_ -2
I
2
4
6
8
10
C
D
.... ~
• ---------'~
o . . . . ~ / ~ ~ - ~ .......... I ""
-2 ;
7 .........
. . . . .
:
"
' ........:t
:
'
'--Z'4--
Fig. 5. Average time courses of BPC angle 7ri for 4 representative normal subjects (1-4) and tilt paradigms A - D (see Fig. 4).
106 for body tilt) and remained essentially constant throughout the measurement episode of 10 min. The average time courses of n i for 4 representative normal subjects (1-4) for paradigms A - D are summarized in Fig. 5. As can be seen, the qualitative shape of the time courses was similar in all subjects, whereas the pre-tilt values and the magnitude of the tilt-induced changes of 7q showed considerable inter-individual differences. The average time constant rH was 6min (S.D. = 4.2 min); the average time constant cawas 5 min (S.D. = 3.4 min). Measurement sequences in the upright position (without tilt) yielded flat time courses with nonsignificant variations. Occasionally subjects were restored to the upright position at the end of a measurement sequence and additional measurement cycles were performed to see whether the tilt-induced changes in n i were reversible. When returned from head or body tilt, ni typically decreased during the first measurement cycle in the upright position and subsequently reached the initial pre-tilt value within the next one or two measurement cycles. When returned from trunk tilts, the return of n i was more gradual, reaching pre-tilt values only after several minutes.
(E) Comparison of left versus right hand pointing It has been reported in a preliminary study 26 that body tilt (right ear down) combined with left hand pointing yielded just mirror-symmetrical results as compared to the right hand/body tilt (left ear down) constellation. Therefore, as a control, we tested pointing performance with the left pointing hand and paradigms A - D for several subjects. It was found that the time courses of n were indeed mirror-symmetrical to those described in part (D) for all paradigms A - D . The individual magnitude of the tilt-induced changes in ni as well as the time constants ZH and Vr corresponded very closely to the values found for the right pointing hand. (F) Postural and visual perceptions All subjects reported postural adaptation during prolonged tilt. When returned to the upright position following sustained body tilt, subjects felt
as though they were being tilted in the opposite direction. Similarly, in the head and trunk paradigm our subjects reported tilt adaptation which diminished gradually with individual time constants ranging from less than one up to several minutes. Although only a single LED was visible at a time, all subjects were able to make approximate estimates of the subjective rotation of the line of visual targets relative to their head. Similar to BPC dichotomy, most subjects also reported a dichotomy in the visual perception of the targets (the contralateral targets were perceived along a line which was almost horizontal relative to the head, whereas the ipsilateral targets were perceived to appear along a line which was rotated relative to the head horizontal). Tilt-induced perceptual rotations of the ipsi- or contralateral half of the target line were mostly coupled with synergistic changes in 7zi and no. DISCUSSION
The present study is based on the novel finding8 that the BPC can typically be subdivided into separate, approximately straight 'branches' for the ipsilateral and the contralateral pointing hemifield. More specifically, the contralateral branch was tyically nearly parallel with the line of visual targets (always horizontal in head coordinates), whereas the ipsilateral BPC branch was rotated by an angle hi, indicating a continuous increase in the vertical error (relative to target line) of blind pointing positions with target eccentricity into the ipsilateral hemifield. This highly reproducible dichotomy of the BPC into two branches is in general agreement with neurophysiological5' ~8.19, behavioural TM, and clinical studies 28, suggesting a hemispherical subdivision of sensori-motor processing in primates. With regard to our suggested scheme for the sequence of coordinate transformations (Fig. 1), it is worth emphasizing that subjects also reported a dichotomy in the perception of the visual targets.
Otolith and neck afferent influence on BPC Stepwise body tilt (left ear down, right hand pointing, Figs. 4 and 5) caused an immediate,
107 stepwise increase in 7q of up to 14 ° with only gradual (if any) subsequent changes. In contrast, body tilt (right ear down) caused a step-like decrease in ni to about zero. As tilt-induced vestibular canal activity should have reached its new steady state following a post-tilt adaptation period of 1 min 36, and no visual spatial reference cues were available, this step-like increase of rri is assumed to be caused by otolith function. This hypothesis rests on neurophysiological data regarding the directional selectivity and static impulse rate modulation of otolith afferents upon adequate stimulation around the roll axis 9'3°. In these terms, the gradual, post-tilt changes in 7~i could reflect long-term otolith adaptation during sustained stimulation 9"12"35. Further evidence supporting our hypothesis of otolith-induced changes is provided by the close correspondence between changes in 7~ and the ocular countertorsion reflex 7. Studies with unilateral labyrinthectomized monkeys 17 and patients with selective otolith dysfunction 4,25 indicate that following lateral tilt, the lower otolith plays a dominant role in the corresponding ocular countertorsion. Preliminary blind pointing studies with unilateral labyrinthine-deficient patients in our own laboratory 13 showed that ni and 7L were typically outside the normal range for lateral tilt towards the deficient ear but normal for tilt in the opposite direction. Stepwise trunk tilt (right shoulder down, right hand pointing, Figs. 4 and 5) caused a gradual decrease in ni (up to 6 ° within 10 min) and to some degree also in no. Stepwise trunk tilt (left shoulder down) caused the reverse effect (gradual increase in 7ti of up to 4 °). As the head was held stationary in an upright position (no otolith stimulation) and the trunk was supported by soft supports (minimizing stimulation of skin receptors), these gradual changes in the BPC are assumed to be caused by neck afferent signals. This hypothesis is compatible with neurophysiological data yielding antagonistic impulse rate modulation of neck afferents during lateral tilt of the trunk relative to the stationary head 6,22,34. The individual time constants (zH and Zr) of the gradual decrease in ni were typically in the range of several minutes, in agreement with the time constants found for the
dynamic changes in the 'subjective vertical' in response to stepwise lateral head tilt24'39). However, the neural time constants of neck afferents are typically in the range of a few seconds 34. Stepwise head tilt (left ear down, right hand pointing, Figs. 4 and 5) led to a transient increase in 7q of up to 9 ° (less than during body tilt, left ear down) which then gradually decreased again, 'finally' arriving at or below the pre-tilt value, reo exhibited only a gradual decrease without a transient response. In contrast, head tilt (right ear down) produced a step-like decrease in 7ti (again less than during body tilt, right ear down) followed by only gradual further changes depending on the subject. Since head tilt involves both otolith stimulation (as during body tilt) and neck stimulation (as during trunk tilt), this PD-type time course would be expected assuming a linear interaction between otolith and neck signals. This notion is supported by additional experiments: combined head tilt (20 o left ear down) and trunk tilt (40 ° left shoulder down) yielded an initial step-like increase in ni (as expected from otolith stimulation) followed by a subsequent further increase (as expected from neck stimulation). Our present findings of separate, dynamic otolith and neck influences on the BPC are in agreement with the neural otolith-neck convergence pattern 3'11.29, the coordination of static limb reflex balance in postural control 2°'21, and the tilt-induced changes of the 'subjective visual vertical '38. Control experiments with the left pointing arm in the reverse tilt paradigms, i.e. body/head tilt (right ear down), trunk tilt (left/right shoulder down), yielded mirror-symmetrical results, thus supporting our hypothesis of a hemifield-specific influence of tilt signals on the neural control of blind pointing movements for each hand. The entire string of blind pointing positions was found to be individually offset relative to the line of the visual targets in vertical as well as in horizontal direction. These shifts of the BPC could be interpreted as offsets of the various postulated neural space maps, see Fig. 1. However, a detailed analysis of this f'mding will be the subject of separate study. On the basis of theoretical considerations, we proposed (see Introduction) that blind pointing
108 involves processing of space coordinates along a chain of various internal neural 'space maps', thus involving intermediate coordinate transformations (Fig. 1) which are assumed to incorporate multimodal tilt information, e.g. from vestibular, visual, proprioceptive receptors. Since blind pointing is performed without visual feedback from the pointing arm (and hand), tilt-induced changes in the BPC possibly reflect the influence of otolith versus neck afferent signals on the neural coordinate transformations, influences which remain obscured (or compensated) under normal closed-loop visual conditions. Our findings suggest that otolith influence on the BPC is asymmetrical, causing significant changes predominantly in the ipsilateral pointing hemifield (rti), whereas the contralateral hemifield (rtc) is rather insensitive to lateral tilt. In contrast, neck afferent stimulation was found to symmetrically influence the pointing hemifields, although ipsilateral changes were typically more prominent than contralateral ones. Both the BPC and the subjective perception of the visual targets underwent marked adaptation upon stepwise change in body position. This is taken as support for the hypothetized existence of an 'internal space map' (Fig. 1) which is an internal representation of space coordinates accessible to both motor program generation as well as to visual perception ~6. Our closed-loop scheme (Fig. 1) of the various coordinate transformations might imply a serial processing of space coordinates through various different neural space maps. We want to emphasize that this sequential arrangement has been chosen for clarity to help the reader to separate and identify the various logical steps the CNS probably has to cope with in the complex task of sensorimotor integration underlying blind pointing. Blind pointing does not necessarily require a serial coordinate transformation process. Rather, it is plausible to assume the existence of a single neural network structure incorporating a continuously updated (via multimodal sensory input) neural space representation, thus making sensorimotor coordinate transformation a parallel process. Our hypothesis of otolith and neck afferent influence on the BPC has been critically assessed in the following ways.
(1) During body tilt, blind pointing movements were the same as those required in upright position but had to cope with an increasing gravitational pull in the direction of the pointing arm. Accordingly, ni might be the result of an overcompensation in motor effort in an attempt to overcome these constraints. We have tested this possibility by supporting the pointing arm with an elastic band from above and found no significant changes in the blind pointing performance. (2) Control measurements with randomized visual target positions insured that our findings were only minimally affected by predictability. (3) The dynamic changes in the BPC following body and trunk tilt might be caused by stimulation of cutaneous receptors in the trunk rather than by otolith and neck signals. However, the following arguments can be made against this possibility. (i) The individual time constants ~. and Vr correspond very closely for a given subject, indicating that the same (neck-) receptor dynamics are involved. This close correspondence also reflects the high individual reproducibility which could hardly be explained by tilt information arising from skin receptors being stimulated in a non-specific manner. (ii) Blind pointing measurements during body tilt (right ear down, right hand pointing) caused an immediate step-like decrease in ni rather than a gradual decrease, although stimulation of skin receptors was similar to that during trunk tilt (right shoulder down). The sensorimotor coordinate transformation paradigm described here requires highly skilled performance and interaction of various sensory, central-processing, and motor activities. Due to the high reproducibility of the BPC and its reversible dependence on the various tilt paradigms, the blind pointing paradigm could therefore be developed into a sensitive indicator for sensorimotor coordinate transformation performance in human patients and normal subjects.
ACKNOWLEDGEMENTS
We thank Ms. A. Offermans for technical assistance. Thanks are also due to Dr. O. Bock for his experimental contributions and discus-
109 sions. Supported by the Deutsche Forschungsgemeinschaft, SFB 200/B10. REFERENCES 1 Bizzi, E., Accornero, N., Chapple, W. and Hogan, N., Arm trajectory formation in monkeys, Exp. Brain Res., 46 (1982) 139-143. 2 Book, O. and Eckmiller, R., Goal-directed arm movements in absence of visual guidance: evidence for amplitude rather than position control, Exp. Brain Res., 62 (1986) 451-458. 3 Boyle, R. and Pompeiano, O., Convergence and interaction of neck and macular vestibular inputs on vestibulospinal neurons, J. Neurophysiol., 45 (1981) 852-868. 4 Brandt, T. and Dieterich, M., Pathological eye-head coordination in roll: tonic ocular tilt reaction in mesencephalic and medullary lesions, Brain, 110 (1987) 649-666. 5 Brinkman, J. and Kuypers, H.G.J.M., Cerebral control of contralateral and ipsilateral arm, hand, and finger movements in the split-brain rhesus monkey, Brain, 96 (1973) 653-674. 6 Denoth, F., Magherini, P.C., Pompeiano, O. and Stanojevic, M., Responses of Purkinje cells of the cerebellar vermis to neck and macular vestibular inputs, Pfliigers Arch., 381 (1979) 87-98. 7 Diamond, S.G. and Markham, C.H., Ocular counterrolling-as an indicator of vestibular otolith function, Neurology, 33 (1983) 1460-1469. 8 Eckmiller, R., Ott, D. and Bock, O., Coordinate transformation of visual target position into blind pointing position depending on the input from otolith organs, 8th International Symposium on Posturography, Amsterdam, 1986, p. 31. 9 Fernandez, C. and Goldberg, J.M., Physiology of peripheral neurons innervating otolith organs of the squirrel monkey. I. Response to static tilts and to long-duration centrifugal force, J. Neurophysiol., 39 (1976) 970-984. 10 Fisk, J.D. and Goodale, M.A., The organization of eye and limb movements during unrestricted reaching to targets in contralateral and ipsilateral visual space, Exp. Brain Res., 60 (1985) 159-178. 11 Fredrickson, J.M., Schwarz, D.W.F. and Kornhuber, H.H., Convergence and interaction of vestibular and deep somatic afferents upon neurons in the vestibular nuclei of the cat, Acta Otolaryngol., 61 (1966) 168-188. 12 Gernandt, B.E., Otolithic influences on extraocular and intraocular muscles, In 5th Symposium on the Role of Vestibular Organs in Space Exploration, SP-314, Washington DC, NASA, 1973. 13 Greven, C., Eckmiller, R. and Ott, D., Unpublished observations, 1986. 14 Guedry, F.E., Orientation of the rotation axis relative to gravity: its influence on nystagmus and the sense of orientation, Acta OtolaryngoL, 60 (1965) 30-48. 15 Howard, J.P. and Templeton, W.B., Human Spatial Orientation, Wiley, London, 1966.
16 Jeannerod, M., How do we direct our actions in space? In A. Hein and M. Jeannerod (Eds.), Spatially Oriented Behaviour, Springer, Berlin, 1983. 17 Krejcova, H., Highstein, S. and Cohen, B., Labyrinthine and extralabyrinthine effects on ocular counter-rolling, Acta Otolaryngol., 72 (1971) 165-171. 18 Lawrence, D.G. and Kuypers, H.G.J.M., The functional organization of the motor system in the monkey. I: the effects of bilateral pyramidal lesions, Brain, 91 (1968) 1-14. 19 Lawrence, D.G. and Kuypers, H.G.J.M., The functional organization of the motor system in the monkey. II: the effects of lesions of the descending brainstem pathways, Brain, 91 (1968) 15-36. 20 Lindsay, K.W., Roberts, T.D.M. and Rosenberg, J.R., Asymmetric tonic labyrinth reflexes and their interaction with neck reflexes in the decerebrate cat, J. Physiol. (Lond.), 261 (1976) 583-601. 21 Magnus, R., KOrperstellung. In Monographien aus dem Gesamtgebiet der Physiologie der Pflanzen und Tiere, Springer Berlin, 1924. 22 Mergner, T., Anastasopoulos, D. and Becker, W., Neural responses to horizontal neck deflection in the group X region of the cat's medullary brainstem, Exp. Brain Res., 45 (1982) 196-206. 23 Mittelstaedt, M., A new solution to the problem of subjective vertical, Naturwissenschaften, 70 (1983) 272-281. 24 Mittelstaedt, M., The effect of visual on extraretinal information about the vertical: suppression or superposition? Proc. XXIII Int. Cong. Psychol., Acapulco, Mexico, Elsevier, Amsterdam, Netherlands, 1984. 25 Nelson, J.R. and Cope, D., The otoliths and the ocular counter-torsion reflex, Arch. Otolaryngol., 94 (1971) 40-50. 26 Ott, D., Eckmitler, R. and Book, O., Transformation of visual target coordinates into blind pointing coordinates depending on the angle of body roll, Neurosci. Lett. Suppl., 22 (1985) 149. 27 Ott, D., Eckmiller, R. and Bock, O., Otolith and neck influence on the blind pointing characteristic, Neurosci. Lett. Suppl., 26 (1986) 327. 28 Perenin, M.T. and Vighetto, A., Optic ataxia: a specific disorder in visuomotor coordination, In A. Hein and M. Jeannerod (Eds.), Spatially Oriented Behaviour, Springer, Berlin, 1983. 29 Pompeiano, O., Manzoni, D., Srivastava, U.C. and Stampacchia, G., Convergence and interaction of neck and macular vestibular inputs on reticulospinal neurons, Neuroscience, 12 (1984) 111-128. 30 Pompeiano, O., Macular input to neurons of the spinoreticulocerebellar pathway, Brain Res., 95 (1975) 351-368. 31 Prablanc, C., Echallier, J.F., Komilis, E. and Jeannerod, M., Optimal response of eye and hand motor systems in pointing at a visual target. I. Spatio-temporal characteristics of eye and hand movements and their relationship when varying the amount of visual information, Biol. Cybern., 35 (1979) 113-124.
110 32 Prablanc, C., Echallier, J.F., Jeannerod, M. and Komilis, E., Optimal response of eye and hand motor systems in pointing at a visual target. II. Static and dynamic visual cues in the control of hand movement, Biol. Cybern., 35 (1979) 183-187. 33 Prablanc, C., Pelisson, D. and Goodale, M.A., Visual control of reaching movements without vision of the limb. I. Role of retinal feedback of target position in guiding the hand, Exp. Brain Res., 62 (1986) 293-302. 34 Rubin, A.M., Liedgren, S.R.C., Milne, A.C., Young, J.A. and Fredrickson, J.M., Vestibular and somatosensory interaction in the cat vestibular nuclei, Pfliigers Arch., 371 (1977) 155-160. 35 Schoen, L., Mikroableitungen einzelner zentraler Ves-
36
37
38
39
tibularisneurone von Knochenfischen bei Statolithenreizen, Z. Vgl. Physiol., 39 (1957) 399-417. Schrader, V., Koenig, E. and Dichgans, J., The effect of lateral head tilt on horizontal post-rotatory nystagmus l and II and the Purkinje effect, Acta Otolaryngol., 100 (1985) 98-105. Taub, E., Goldberg, I.A. and Taub, P., Deafferentation in monkeys: pointing at a target without visual feedback, Exp. Neurol., 46 (1975) 178-186. Wade, N.J., Visual orientation during and after lateral head, body, and trunk tilt, Percept. Psychophys., 3 (1968) 215-219. Wade, N.J., Effect of prolonged tilt on visual orientation, Q. J. Exp. Psychol., 22 (1970) 423-439.