The Effect of Central Visual Impairment on Manual Prehension When Tasked with Transporting-to-Place an Object Accurately to a New Location

Abstract

Purpose.: To investigate whether differences exist between visually impaired and normal vision participants when completing a visually guided manual prehension task of transporting-to-place an object accurately to a new location.

Methods.: Participants with central visual impairment (n = 17) and normal vision (n = 10) were tasked with reaching-to-grasp an object and then transporting-to-place it accurately to a different (predefined) location. Analysis assessed movements of the object transport phase and error in object placement. Correlation analysis assessed the association between the movements and mean loss in the central 5°, 10°, and 20° of participant's integrated binocular visual field (VF), visual acuity (VA), and contrast sensitivity (CS).

Results.: Compared to normal vision participants, visually impaired participants took significantly longer to complete the movement and exhibited increased deceleration time, greater number of velocity corrections, and more errors in object placement (P < 0.05). Findings demonstrated that participants with central visual impairment were more cautious when executing the movement, which contributed to uncertainty regarding end-position location. A higher number of movement indices were correlated with loss in the central 5° compared to the 10° and 20° of the VF. VA and CS were significantly correlated with a number of movement variables.

Conclusions.: Despite the increased time taken by visually impaired participants to complete the task, they were still less accurate than normals when placing the object. Decrements in movement control were most associated with loss in the central 5° when compared to the central 10° and 20° of the VF, which was evidenced in the deceleration/“online” component of the movement.

Introduction

With the important role that vision plays in planning and executing manual prehension/reach-and-grasp, ^1–3 it is not surprising that visual impairment significantly influences the ability to complete such a task. Indeed, decrements in reach-and-grasp performance have previously been highlighted through simulating both peripheral ^4–7 and central ⁶ visual impairment in young normal vision adults, and in actual visual impairment in older adults with peripheral and central impairment from eye diseases such as glaucoma ⁸ and age-related macular degeneration. ^9–11  

Since we live in an aging population ^12–14 and visual impairment is linked with increased age, ^15,16 it is important that we continue, through ongoing investigation, to increase our understanding of the effects that visual impairment has on daily function. A greater understanding will facilitate the development of increasingly effective rehabilitation techniques. While research that simulates visual impairment in young normal vision adults provides some important insights into the functional limitations associated with visual impairment, ^4–7 these findings are limited when drawing inferences about the visually impaired population as it does not account for long-term visuomotor adaptation to vision loss ⁹ or the effects of age on manual prehension. ¹⁷ Research investigating the effects of actual visual impairment on daily function has highlighted that when patients with peripheral or central visual impairment complete a reach-and-grasp task, they exhibit delayed onset and longer movement times, ^8–11 lower peak velocities, and impaired grip scaling (central visual impairment only^9–11) compared to age-matched adults with normal vision. These aforementioned studies all measured participants' ability to reach-and-grasp an object, whereas, in daily life, we are required to perform a variety of other manual prehension tasks such as grasping objects and then transporting-to-place them accurately to a different location; how this affects individuals with central visual impairment has not been previously investigated. 

Visually impaired individuals self-report difficulties for tasks that require some element of object transport and placement, for example, setting the table, ^18,19 putting away groceries, ¹⁸ and organizing objects. ²⁰ To further increase our understanding of the impact visual impairment has on daily function and to quantify the difficulties associated with object transport and placement, the current study recruited participants with central visual impairment who completed a novel task of reaching-to-grasp an object and then transporting-to-place it accurately to a different (predefined) location. Prehensile movements were objectively assessed using a Vicon motion analysis system, a sophisticated analysis tool that allows precise measurements during an activity to be recorded. We also investigated the relationship between visual field (VF) loss (and other aspects of visual function) on movement kinematics and accuracy of object placement. The current research extends upon the previous work conducted by Pardhan et al. ^9,10 and Timberlake et al., ¹¹ through investigating how participants with central visual impairment transport and place an object accurately at a different (predefined) location rather than just reaching-to-grasp an object. The reach and grasp paradigms used previously have only required participants to reach and grasp objects placed directly in front of them. This research extends the potential area of VF that would be used from that in the traditional reach and grasp task. Moreover, this study examines parameters that have not been examined before in studies on reaching and grasping in visual impairment.^9–11 Since participants were required to locate and pick up the object and then to also locate the end position through visual search, there is arguably a greater visual demand placed on the participant compared to the previous reach and grasp tasks. This reflects a more realistic task found in the home/kitchen. 

Methods

Participants

Seventeen participants with central visual impairment (age 82 ± 10 years) and 10 participants with normal vision (age 77 ± 5 years) participated in the study. The mean ages between groups were not significantly different (t-test, P = 0.11). Health and physical fitness of all participants were assessed through a self-report questionnaire. Participants were excluded if they reported any history of neurological, musculoskeletal disorders that could affect prehension performance, had insulin-dependent diabetes, or failed to score the minimum requirement to pass the Mini Mental State Examination (MMSE ²¹ ). Participants were also given a complete eye examination prior to their inclusion in the study. This included a cover test and a motility test. Participants were excluded from the study if they had amblyopia or strabismus or if their motility (which extended to all directions of gaze) was not considered normal or was decompensated. This ensured that the binocular eye movements of all participants were normal. Based on the above criteria, one participant (visually impaired) was excluded from the study. Hand function was also assessed by a human movement scientist, and each participant was deemed to have normal range of motion. Visual examination of all participants was provided by a consultant ophthalmologist. All participants with central VF loss had diagnosed bilateral macular problems; 16 participants were diagnosed with age-related macular degeneration and 1 with a macular hole. The Tenets of the Declaration of Helsinki were observed, and the experiment gained approval from the University's Ethical Committee. Written informed consent was obtained from each participant prior to undertaking the study. 

Visual Assessments

Distance visual acuity (VA) and contrast sensitivity (CS) were measured both monocularly and binocularly. However, since all prehension movements were completed with both eyes, only binocular scores are reported. VA was measured using a Bailey-Lovie LogMAR chart at a working distance of 4 m, using a letter-by-letter scoring system (0.02 LogMAR). If participants were unable to read the largest letters on the LogMAR chart at 4 m, shorter distances were used and the score adjusted accordingly. CS was measured using the Pelli-Robson chart at 1 m and scored by letter (0.05 log units). Stereopsis was measured using the Frisby stereo test, initially at a distance of 60 cm. All visual measurements were carried out by an optometrist under normal room illumination (531 lux) using the best-corrected spectacle prescription for that distance as determined by subjective refraction. The mean binocular VA scores for visually impaired and normal vision group were 0.81 ± 0.42 LogMAR and −0.04 ± 0.09 LogMAR, respectively. CS scores for visually impaired and normal vision group were 1.06 ± 0.27 log and 1.71 ± 0.10 log, respectively. From the 17 visually impaired participants recruited, only 2 had measurable stereoacuity (both 340 seconds of arc). The normal vision group recorded 59 ± 16 seconds of arc (see Table A1 for individual participant characteristics). 

Visual field assessment was conducted using Humphrey Field Analyzer (Carl Zeiss Meditec, Inc., Dublin, CA) SITA-Standard 24-2 threshold test. All participants were tested monocularly wearing their best near correction. Care was taken to ensure that the blind spot was located as accurately as possible in the visually impaired participants, which subsequently resulted in the correct retinal locations being used for the integrated binocular VF. Binocular VFs were calculated based on the more sensitive of the two VF locations for each eye (i.e., the “best location” model ²² ). 

To ascertain whether specific areas of the VF were correlated to manual prehension performance, based upon the integrated binocular VF plot, the mean threshold ^23,24 was defined for the central 5°, central 10°, and central 20° for each participant, which identified the central 4, 16, and 36 test points respectively (see Fig. 1 below). The assessment of the different extents of the binocular VF in this manner was based on published work that investigated the association between the VF and the assessment of vision ²³ and perceived difficulty with performing daily living tasks. ²⁴  

Figure 1.

 

Integrated binocular visual field plot for a visually normal participant with illustration of the central 5, 10, and 20 degrees grids overlaid.

Protocol

Using the thumb and index finger to adopt a precision grip, participants were tasked with reaching-to-grasp a white cylindrical object either “large” or “small” (80 mm × 100 mm, 30 mm × 100 mm, respectively; diameter × height) that was initially placed in start position “near” or “far” (360 mm or 560 mm, respectively) from them along their midline and then transporting-to-place it accurately to a different (predefined) location either “end position near” or “end position far” (150 mm or 350 mm, respectively) lateral on the ipsilateral side of dominant hand (see Fig. 2). The angles of eccentricity between the start and end positions were calculated from the average vertical eye position, which was approximately ∼ 560 mm from the table top. This angle was calculated based on the distance (hypotenuse) of the eyes from each start position and end position of the object, taking into account the distance at which the object was placed (laterally from the participant) and the vertical height of the eyes from the table top. It was necessary to use the hypotenuse to calculate the angle of eccentricity, as the eyes were not positioned at the height of the table top during the experiment. Based on a hypotenuse distance of 666 mm from the center of the two eyes to start position “near” and 792 mm to start position “far,” the following angles of eccentricity were calculated. 

Figure 2.

 

Illustration of the experimental paradigm. Participants were required to pick up the object (placed in start position near or far) and transport and place the object at the end position (near or far), which was on the same lateral plane from where the object was initially grasped. Note that the illustration is of a right-handed participant.

The angle at eye level (with the participant looking straight ahead) from start position “near” (360 mm anterior of subject) to end position “near” (150 mm lateral of initial object start position) was 13° eccentricity. The angle with same start position “near” to end position “far” (350 mm lateral of initial object start position) was 28° eccentricity. The angle of eccentricity for picking up the object from start position “far” (560 mm anterior of subject) and transporting to end positions “near” and “far” resulted in 11° and 24° eccentricity, respectively. Pilot work also determined that placing the object at end position “far” (350 mm) was within the region in which participants could comfortably transport the object without having to assume an unnatural/uncomfortable position. Care was taken to ensure that each participant was positioned at approximately the same height from the table (for consistency of eccentric angle) prior to starting the movement. 

Participants were seated comfortably on a chair in front of a flat table (120 cm × 80 cm) that was covered with a black cloth. Participants placed their dominant hand (determined by the Edinburgh Inventory ²⁵ ) at the edge of the table at a predefined position, located 200 mm from the participant's midline on the ipsilateral side of their dominant hand (position approximately in line with the shoulder). Once participants had positioned their hand, they were instructed to look straight ahead and close their eyes; this was done to standardize the initial position of the hand and to enable the randomization of the object size, object start position, and end-point location. Once participants had closed their eyes, the object (either small or large) was positioned in the initial start position (either near or far) and they were instructed that on the “go” command (which coincided with opening their eyes) they were to pick up the object and transport it to “near end position” or “far end position” location. Participants were instructed to complete the task as they would when completing similar everyday activities. Participants did not receive any information prior to the start of the movement, and all parameters (object size and object start and end position) were fully randomized. Once participants were given the command to start the movement, they opened their eyes, grasped the object, and repositioned it in the required end position. The location of each end position was marked with a white “+” (20 mm × 20 mm). Participants were instructed to reposition the object as accurately as possible, while being made aware that they were being timed. All trials were completed under binocular viewing condition, with participants optimally corrected for a reach of 46 cm, the average distance between the two object start positions. Sufficient training was given to ensure that all participants were familiar with the protocol prior to the experiment. 

Prior to the start of the experiment, participants completed a pretest in which they grasped the large and small object at each initial start position and placed the object in each end position, resulting in a total of eight trials being collected. These trials were different from the main test in that participants were allowed to reposition/move the object after it had been placed on the table if they thought they could be more accurate. In the main test this was not permitted; once the object had been repositioned on the table it was not allowed to be moved. During the pretest trials, participants were instructed to take as much time as they required to place the object as accurately as possible, with emphasis placed on accuracy rather than speed. The “x” and “y” coordinates from each object placement were recorded through the Vicon motion capture system, and the accuracy (in terms of errors) of each subsequent object placement in the main experiment was compared to the relevant pretest trial. 

Each object size (small or large), object start position (near or far), and object end position (near end position or far end position) was repeated three times in a fully random order, which resulted in a total of 24 trials being completed. 

Kinematic data were collected (at 100 Hz) using a six-camera 3-D motion capture system (Vicon 460, Oxford Metrics Ltd., Oxford, UK). Additional information pertaining to the capture system can be found elsewhere. ⁹ Retro-reflective spherical markers were attached at the following key anatomical locations (placed either directly on the skin or clothing): distal border of the thumbnail and index fingernail, styloid process on the radial side of the wrist, sternum, and the anteriolateral and posteriolateral aspects of the head. An additional marker was attached at the top of each object at the center. Marker trajectory data were filtered using the cross-validatory quintic spline smoothing routine with “smoothing” options set at a predicted mean square error value of 10 and processed using the PlugIn-Gait software (Oxford Metrics Ltd.). 

Data Analysis

The main focus of this paper was to assess movements from the period when the object was initially grasped to when it was repositioned in the new location. Analysis focused on determining variables that we and others have previously found important for manual prehension. ^{2,9–11,26,27} The variables listed below were assessed; velocity profiles were all calculated from the wrist marker. 

1.  
   Movement time: The time from when the resultant velocity of the object was first displaced by greater than 10 mm/s for five consecutive frames to when it remained less than 10 mm/s for five consecutive frames. Although previous research has used a range of criteria to define movement onset (e.g., 50 mm/s for five consecutive frames ^2,28 ), analysis of our data highlighted that 10 mm/s was the most appropriate indicator of movement.
2.  
   Peak lateral (side to side) head movement: Peak lateral head movement calculated relative to the head angle when looking straight ahead. This was assessed to determine whether visually impaired patients moved their head differently compared to normal vision participants when placing the object at the end-point location.
3.  
   Peak velocity: Maximum velocity of the wrist for both lateral and vertical directions during the movement. These indices indicate the feedforward or planning component of the movement. ²
4.  
   Time to peak velocity: Time from movement onset to both peak lateral and vertical velocity.
5.  
   Deceleration time: Time from both peak lateral and vertical velocity to end of the movement; indicative of the online control of the movement. ^26,27
6.  
   Peak height of the wrist: Peak vertical height of the wrist during the movement; indicative of the feedforward component of the movement. ²
7.  
   Velocity corrections: Additional movements in the velocity profile after peak velocity to end of the movement for both lateral and vertical directions. ²⁷
8.  
   Error in object placement: Error in object placement was determined by calculating the difference in resultant “x” and “y” coordinates at the end of each trial to the appropriate pretest trial. Analyzing the error in object placement would provide an overview of the efficiency of the movement. ²⁶

Statistical Analysis

Normal distribution of the data was confirmed using the Kolmogorov-Smirnov test (P > 0.05). Processed data were analyzed using separate ×2 group (visually impaired, normal) ×2 object size (small, large), ×2 initial object position (near or far) ×2 end-point location (near end point, far end point) mixed between groups ANOVA, with repeated measures on each factor. Level of significance was accepted at P < 0.05, and post hoc analyses where appropriate were performed using Tukey's HSD. 

To determine the significance of association between specific aspects of the movement and visual function (VA, CS, central 5°, 10°, and 20° of the VF), correlation analysis was conducted. Movement data were averaged across all conditions (object size, object start and end position). VA, CS, and central 5° of the VF were not normally distributed and were analyzed using Spearman's rank order correlation coefficient; 10° and 20° of the VF were normally distributed and were analyzed using Pearson's correlation coefficient. 

Results

Since previous research has already examined the initial reach-and-grasp among patients with central visual impairment,^9–11 we do not report this part of the movement in the paper; however, to highlight the consistency of our work to previous reach-and-grasp research, ^9–11 these results can be found in Table A2. These data support and confirm results of earlier work. ^9–11  

From the 810 trials completed within the study, the object was only knocked over on four occasions (all by visually impaired participants when initially being grasped). These trials were discarded from the analysis. None of the objects fell over when being repositioned at the end position or were placed in the wrong end position. Some participants did, however, exhibit error in the precise placement of the object (see error in object placement variable). 

Movement Time.

Movement time was significantly affected by group (P < 0.001) and end position (P < 0.001, see Table 1), and there was also a significant group-by-end-position interaction (P = 0.01). Post hoc testing showed that visually impaired participants took significantly longer to complete the movement compared to normal vision participants when transporting the object to both end positions; however, the increase in movement time when transporting the object to far end position compared to near end position was significantly longer in the visually impaired group only (see Fig. 3, left). Movement time was unaffected by object size or object start position (P > 0.05). 

Table 1.

 

Movement Kinematics; Group Mean (±SD) across Visual Group and End-point Location

	Visually Impaired		Normal		ANOVA
	Near End Position	Far End Position	Near End Position	Far End Position
Movement time (s)	1.78 (0.51)	2.25 (0.53)	1.29 (0.20)	1.5 (0.26)	G*, EP*, (g-ep)
Peak lateral head movement (deg)	1.35 (0.76)	1.35 (0.79)	0.58 (0.54)	0.60 (0.57)	G, OBJ*, IP*, EP
Lateral deceleration time (s)	1.23 (0.46)	1.60 (0.49)	0.85 (0.18)	0.97 (0.23)	G, EP*, (g-ep)
Vertical deceleration time (s)	1.35 (0.50)	1.78 (0.52)	0.93 (0.21)	1.07 (0.27)	G, EP
Peak vertical height of the wrist (mm)	124 (39)	136 (42)	79 (21)	96 (28)	G, IP, EP*
Lateral velocity corrections (n)	40 (23)	48 (23)	23 (9)	21 (10)	G
Error in object placement (mm)	11 (8)	11 (10)	5 (2)	6 (6)	G, OBJ

 Table shows data averaged across initial start position (IP) and object size (OBJ) to illustrate the effects of vision group (G) and end-position location (EP). Factors found to be significant are shown by capital letter (P < 0.05) and asterisks (P < 0.001), and interactions between factors are denoted by lowercase letter in parentheses.

Figure 3.

 

Group mean (±SD) (left) movement time and (right) lateral deceleration time for visually impaired (VI) and normal vision (Norm) participants when transporting the object to near and far end position.

Peak Lateral Head Movement.

Peak head movement was significantly increased in the visually impaired group (P = 0.007) when grasping the smaller object (P < 0.001), when the object was initially placed in the far start position (P < 0.001), and when transported to far end position (P = 0.03, see Table 1). There were no significant interactions. 

Peak Lateral Velocity.

Peak lateral velocity was significantly higher when the object was transported to far end position compared to near end position (P < 0.001); there were no other significant effects or interactions (P > 0.05). 

Time to Peak Lateral Velocity.

Time to peak lateral velocity occurred significantly later in the visually impaired group compared to normal vision participants (P < 0.05) and when the object was being transported to far end position compared to near end position (P < 0.001, see Table A3). There were no significant interactions. 

Lateral Deceleration Time.

Lateral deceleration time was significantly affected by group (P = 0.003) and end position (P < 0.001). It was unaffected by size or initial object start position. There was a significant group-by-end-position interaction (P = 0.005), which indicated that lateral deceleration time was increased in the visually impaired compared to normal vision participants when transporting to far end position only. There was also a significant increase in lateral deceleration time when transporting to far end position compared to near end position in the visually impaired group only (see Fig. 3, right). 

Peak Vertical Velocity.

Peak vertical velocity was unaffected by group, initial object start position, or object size. There was, however, a significant effect on end position (P = 0.002) and group-by-end-position interaction (P = 0.04). Post hoc analysis showed that there was a significant increase in peak vertical velocity in the normal vision compared to the visually impaired group when transporting the object to far end position compared to near end position (see Fig. 4). 

Figure 4.

 

Group mean (±SE) maximum vertical velocity for visually impaired (VI) and normal vision (Norm) participants when transporting the object to near and far end position. A significant group-by-end-position interaction effect was observed (P = 0.04).

Time to Peak Vertical Velocity.

Time to peak vertical velocity occurred significantly later among visually impaired compared to normal vision participants (P < 0.05); there were no other main effects or interactions (P > 0.05). 

Vertical Deceleration Time.

Vertical deceleration time was significantly longer among the visually impaired group compared to normal vision participants (P < 0.05) and smaller when transporting the object to near end position compared to far end position (P < 0.007); there were no other main effects or interactions (P > 0.05). 

Peak Vertical Height of the Wrist.

Peak vertical height of the wrist was significantly higher among the visually impaired group (P = 0.003, see Fig. 5) when the object was initially placed in the far start position (P = 0.01) and when transported to far end position (P < 0.001); there were no other main effects or interactions (P > 0.05). 

Figure 5.

 

Minimum (min), maximum (max), and average vertical hand trajectory for visually impaired (VI, shaded symbols) and normal vision (Norm, open symbols) participants. Exemplar data are presented for the small object transported to far end position. ∼“0” coordinate on the x axis represents start position and ∼350 mm represents end of the movement. Note that the figure only includes right-handed participants. There was a significant difference in vertical movement between groups (P = 0.003).

Lateral Velocity Corrections.

Visually impaired patients exhibited a significantly greater number of lateral velocity corrections compared to normal vision participants (P = 0.04, see Table 1); there were no other main effects or interactions (P > 0.05). 

Vertical Velocity Corrections.

There were significant effects of group (P = 0.012) and end position (P < 0.001); however, object size and object start position had no effect. There was a significant group-by-end-position interaction (P = 0.025), which indicated that there was a significant increase in the number of vertical velocity corrections among the visually impaired group when transporting the object to far end position compared to near end position and also compared to normal vision participants when transporting the object to far end position only (see Fig. 6). 

Figure 6.

 

Group mean (±SD) number of vertical velocity corrections for visually impaired (VI) and normal vision (Norm) participants when transporting the object to near and far end position. A significant group-by-end-position interaction was observed (P = 0.025).

Error in Object Placement.

Resultant error in object placement was significantly greater among visually impaired compared to normal vision participants (see Fig. 7). Both groups exhibited greater error when repositioning the large compared to small object (P < 0.04); there were no other main effects or interactions (P > 0.05). 

Figure 7.

 

Resultant error in object placement in visually impaired (VI, shaded symbols) and normal vision participants (Norm, open symbols) when placing the (a) small (S) and (b) large (L) objects. Absolute values are presented. Note that “0,0” would represent placing the object in the same location as the pretest trial. A significant effect was observed for group and object size (P < 0.04).

Correlation Analysis

CS and VA were significantly correlated with the highest number of dependent variables (11 variables each, see Table 2 below and Table A4). Central 5° of the VF was correlated with nine variables, and central 10° and 20° of the VF correlated with four and five variables, respectively. See Table 2 and Table A4 for specific details. 

Table 2.

 

Correlation Coefficients between Movement Kinematics and Visual Function

	VA			CS			Central 5°			Central 10°			Central 20°
	r	r 2	P	r	r 2	P	r	r 2	P	r	r 2	P	r	r 2	P
Movement time	−.522	0.27	.005	−.735	0.54	.000	−.631	0.40	.000	−.477	0.23	.012	−.448	0.20	.019
Peak lateral head movement	.518	0.27	.007	−.500	0.25	.009	−.269	0.07	.183	−.346	0.12	.083	−.398	0.16	.044
Lateral deceleration time	.446	0.20	.020	−.700	0.49	.000	−.467	0.22	.014	−.378	0.14	.052	−.370	0.14	.057
Lateral velocity corrections	.472	0.22	.013	−.674	0.45	.000	−.507	0.26	.007	−.271	0.07	.171	−.293	0.09	.138
Vertical velocity corrections	.416	0.17	.031	−.638	0.41	.000	−.409	0.17	.034	−.313	0.10	.112	−.314	0.10	.111
Error in object placement	.701	0.49	.000	−.664	0.44	.000	−.508	0.26	.007	−.272	0.07	.169	−.296	0.09	.134

 Significant correlations are denoted in bold.

Discussion

In daily life we regularly perform a variety of manual prehension tasks that not only require us to reach-to-grasp an object but also to transport an object and place it accurately at a different location. People with visual impairment have self-reported difficulties for tasks that involve some element of object transport and placement, for example, setting the table, ^18,19 putting away groceries, ¹⁸ and organizing objects. ²⁰ To further increase our understanding of the impact visual impairment has on daily function and the difficulties associated with object transport and placement, participants with central visual impairment were tasked with reaching-to-grasp an object and then transporting-to-place it accurately to a different (predefined) location. Analysis focused on the movement kinematics of the object transport phase and accuracy of object placement. The study also investigated the relationship between size/severity of VF loss (among other aspects of visual function) in each participant and movement kinematics. Results from the study highlight that, compared to participants with normal vision, participants with central visual impairment took significantly longer to complete the movement and were less accurate in placing the object at the predefined location. Results also highlight that more kinematic indices were correlated with loss in the central 5° of the VF compared to loss in 10° and 20° of the VF. This is likely attributable to VF loss being most severe in the central 5° of the VF among our visual impaired participants. CS and VA were also strongly associated with a number of kinematic variables. Findings from the initial reach and grasp are reported in Table A2, which confirms results of previous studies. ^9–11 Indeed, compared to normal vision participants, visually impaired participants took significantly longer to initiate and complete the initial reach-and-grasp movement and had significantly longer deceleration time and time after maximum grip aperture. 

When completing precision movements, a speed-accuracy trade-off exists (termed Fitt's law), whereby increasing the time taken to complete the task facilitates more accurate results. ²⁹ However, in the current study, despite the increased time taken by visually impaired participants to complete the precision task (see Fig. 3, left), they were still less accurate than normal vision participants when placing the object (see Fig. 7). The increased error in object placement among visually impaired participants is likely attributable to the degradation of their central VF. Indeed, participants with normal vision would have likely used visual information gained from the fovea (the area of the eye that provides the highest amount of resolution information) to precisely place the object. However, since visually impaired participants had a macular scotoma that either occluded or significantly degraded visual information available from the fovea, when placing the object they would have either been required to rely on impoverished visual information from the fovea or (more likely) use a more eccentric fixation (commonly referred to as a preferred retinal locus ^30–32 ), which has a poorer visual resolution and reduced fixation stability than the fovea. ¹¹ Reductions in visually impaired patients' VA and CS also likely contributed to the increased error in object placement. Both groups were less accurate when repositioning the large compared to small object. A larger object with a greater diameter would create more uncertainty around the end-position location immediately prior to being placed compared to a smaller object that covers less area. Interestingly, however, this uncertainty did not result in an increased number of vertical or lateral velocity corrections prior to placing the object. 

Not only were visually impaired participants less accurate when repositioning the object, they also showed greater uncertainty regarding location of end position, as evidenced by increased lateral and vertical deceleration time and a greater number of lateral and vertical velocity corrections (vertical velocity corrections only significant between groups when transporting to far end position; Fig. 6). By increasing the deceleration phase, this provided a longer opportunity during the final portion of the movement to make online corrective adjustments ³³ in an attempt to minimize the error between actual and required object end position. Visually impaired participants exhibited a significantly greater number of velocity corrections compared to normal vision participants during the deceleration phase, indicating that they were uncertain regarding precise end position. This type of movement behavior has been similarly demonstrated during reach-to-grasp in participants with central ^9–11 and peripheral ⁸ visual impairment. 

In the present study, compared to normal vision participants, visually impaired participants lifted their wrist/object significantly higher during the transport phase of the movement (Fig. 5). Similar adaptations in movement trajectory are shown in patients with age-related macular degeneration who produce a less “direct” reach when reaching-to-grasp an object. ¹¹ The reason for such adaptation is unclear; however, one explanation is that transporting the hand using a less direct trajectory ensures that it is not obscured by the central scotoma. ¹¹ In the current study it is also possible that degradation of the central VF caused participants to lift their hand higher (and by implication use a less direct trajectory) as the result of a safety strategy to reduce the risk of the wrist (and object) contacting the table and causing an injury; this is supported by the correlation of peak vertical height with mean loss in the central 5° of the VF (Table 2). This adaptation in movement control would reduce the occurrence of the object colliding with the table or indeed other objects (if placed on the table) when transporting the object. 

Visually impaired participants exhibited significantly greater peak lateral head movements compared to normal vision participants (see Table 1). This result could be a function of two adaptations; the increase in lateral head movement may have simply been a consequence of using their preferred retinal locus and ensuring that their scotoma did not occlude the end position for eventual object placement. It is also possible that this was a behavioral adaptation to increase the retinal motion information derived from movements of the head, such as depth and direction information from motion parallax ³⁴ and optic flow, ³⁵ to help control the movement. This is supported by Marotta et al., ³⁶ who highlight that, when normal vision participants were required to reach-to-grasp an object with monocular vision, they exhibited fewer corrective movements when they were allowed to move their head compared to it being fixed. 

In the current study, visually impaired patients had either no or minimal stereopsis (see Table A1), which would have resulted in impaired depth perception. Lack of depth cues has been shown to impact reaching and grasping tasks in young participants. ^3,26,27 On the other hand, Timberlake et al. ¹¹ suggest no significant differences between reach and grasp performance in monocular and binocular conditions in their older normals (except for maximum velocity at 20-cm reach) or in their scotoma patients. This lack of difference between monocular and binocular behavior in their normal vision participants may be because these participants were older and were more likely to have unequal monocular acuities, which would impact on their overall binocular acuity. If differences exist between the two eyes, then it is likely that depth perception cues are minimal. Inspection of the visual acuity data of their normal vision subjects demonstrates differences in monocular visual acuities in 7/10 subjects while 3/10 have equal acuity in both eyes. While Timberlake et al. ¹¹ report that all their subjects “could see the stereoscopic images in the Titmus Fly Test,” it is unclear whether differences in the degree of stereopsis exist within these patients. In the current study, we would predict little difference between binocular and monocular performance in our visually impaired patients, as we expect the binocular vision to be driven by the good eye and minimal depth cues would be available. Since our normal vision subjects had similar acuities in each eye and had relatively good stereopsis, differences in monocular behavior compared to binocular may well have existed in our normal subjects. Whether this would be significantly different for the age group of the participants in our study warrants further investigation. 

Important interactional effects were also observed between vision group and end position. While visually impaired participants exhibited significantly longer movement times when transporting the object to far end position compared to near end position, there was no difference for normal vision participants (Fig. 3, left). The increased movement time is attributable to an increased lateral deceleration time when transporting to far end position compared to near end position (Fig. 3, right), which remained unchanged among normal vision participants. Impaired performance showed by visually impaired participants was therefore more evident when transporting the object longer distances. 

It is possible that some of our results may be explained by an increased “pathway time or saccadic time” of eye movements in patients with visual impairment. Despite confirming that binocular eye movements of visually impaired patients were normal, there may well be more errors in their eye movement trajectory when completing the task. However, this was not measured, as it would go beyond the scope of the study, require additional hardware (i.e., eye tracker or scanning laser ophthalmoscope), and, more importantly, might pose additional constraints on the participants when carrying out this task (through wearing an eye tracker), which would prevent this task from being habitual. Exploring whether the differences between the two groups are due to eye movement differences (which may result from differences in their visual function) is something that can be explored in a follow-up study. Moreover, a previous study ³⁶ exploring eye-hand coordination suggests that although the performance of patients with visual impairment is significantly worse than normal subjects, the fundamental features of eye-hand coordination remain the same. 

Correlation Analysis

This study investigated whether specific areas of the VF were correlated with manual prehension performance and whether severity/extent of VF loss was associated with decrements in performance. Results highlight that mean loss in the central area of the VF (i.e., central 5°) is associated with a greater number of movement indices compared to the mean loss in the larger areas of the VF (central 10° and 20°, see Table 2 and Table A4). This is likely attributable to the fact that VF loss was most severe in the central 5° of the VF compared to central 10° and 20° of the VF among our visually impaired patients. 

Interestingly, while general movement profiles (e.g., movement time) were correlated with mean loss in all sizes of the VF, the deceleration phase of the movement was only correlated with mean loss in the central 5° of the VF. Indeed, the number of velocity corrections during the final part of the movement and error in object placement were significantly associated with loss in the central 5° but not loss in the central 10° or 20° of the VF. Since the fovea provides the highest amount of resolution information, loss in the central field will impact the ability to “fine-tune” the latter part of the movement. This is similarly reported by Timberlake et al. ³⁷ who found that a patient with a large central scotoma was less accurate when completing a range of manual prehension tasks despite increasing the online component of the movement through fixating earlier and for longer on the end point compared to a visually normal participant. While subtle differences in fixation behavior were reported between scotoma and normal participant, the fundamental features of eye-hand coordination remained the same in that the “eyes led the hand/fingertip” during the movement. While only the central 5° was associated with the deceleration component of the movement, only the larger 20° VF was associated with peak head movement (see Table 2). As one would expect, the larger area of the VF would be required to take into account the overall change in the visual scene as the individual moved their head. 

Results also support previous research highlighting that degraded VA and CS are significantly associated with decrements in movement control. ^9,11,38,39 However, unlike previous research, ^9,40,41 we did not find that a higher number of movement variables were associated with CS than VA, rather there was an equal association (see Table 2 and Table A4). This may be attributed to the different movement indices analyzed and/or different task used. Results also showed that both VA and CS were associated with a greater number of movement indices than the central 5°, 10°, or 20° of the VF, which suggested that VA and CS are more important for assessing visual dysfunction in macular patients than the VF. Clearly further research is needed to confirm this. 

Conclusion

When participants with central visual impairment transported an object and placed it accurately in a new location, they took significantly longer to complete the movement and were less accurate in placing the object compared to participants with normal vision. This was likely attributable to uncertainty regarding location of end position as a result of their scotoma. Results also highlighted that decrements in movement control were associated more with the central 5° of the VF when compared to central 10° and 20°, which was evidenced in the deceleration component of the movement. 

Future work should continue investigating the effect that visual impairment has on daily living by considering tasks which are different from those highlighted throughout this paper. Indeed it is only through ongoing investigation that it will be possible to gain a holistic understanding of the functional disability caused through visual impairment. 

Acknowledgments

The authors thank Daryl Tabrett for conducting the visual assessments on the participants and Stephanie Wong for her assistance during data collection and processing. 

Footnotes

 Disclosure: M.A. Timmis, None; S. Pardhan, None

Appendix A

Table A1.

 

Data for VI and Norm Participants

	Age	CS	VA	Stereopsis	Central 5°	Central 10°	Central 20°
VI 1	56	1.25	1.16	NA	30	30	30
VI 2	67	1.15	0.5	NA	14	21	23
VI 3	82	1.2	0.44	NA	23	24	23
VI 4	83	0.85	1.14	NA	14	12	14
VI 5	83	1.4	0.42	NA	30	30	30
VI 6	92	1.25	−0.04	NA	22	25	25
VI 7	90	0.75	0.9	340	27	28	27
VI 8	95	1.05	0.84	NA	11	16	19
VI 9	80	0.5	1.32	NA	7	10	9
VI 10	89	0.95	1.2	NA	19	21	22
VI 11	81	1.05	0.64	NA	25	26	25
VI 12	95	0.85	1.12	NA	16	18	20
VI 13	79	1.05	1.14	NA	10	11	10
VI 14	89	1.05	1.1	NA	26	23	24
VI 15	82	1.15	1.12	NA	23	22	20
VI 16	70	1.7	0	340	33	32	31
VI 17	81	0.9	0.8	NA	18	23	24
Norm 1	76	1.8	−0.2	40	32	32	31
Norm 2	86	1.8	0.02	40	33	31	31
Norm 3	80	1.7	0.02	75	29	28	27
Norm 4	77	1.7	0.04	55	30	27	26
Norm 5	74	1.9	−0.10	55	32	31	30
Norm 6	72	1.65	−0.08	70	31	30	29
Norm 7	82	1.55	0.00	40	31	30	29
Norm 8	79	1.65	−0.08	80	32	30	28
Norm 9	74	1.65	0.10	80	29	28	28
Norm 10	72	1.7	−0.10	55	30	27	27
Average VI	82	1.06	0.81	NA	20	22	22
Average Norm	77	1.71	−0.04	59	31	29	29

 Average scores of all the data points for binocular central 5, 10, and 20 degrees of the visual field (scored in dB). VI, visually impaired; Norm, normal vision.

Table A2.

 

Initial Reach-and-Grasp Performance between Visually Impaired and Normal Vision Participants

Reach-to-Grasp Parameter	Mean (±SD)		ANOVA
	Normal	VI
Movement initiation (s)	0.52 (0.22)	0.82 (0.33)	G*
Movement time (s)	1.22 (0.28)	1.78 (0.51)	G†
Peak velocity (mm/s)	735 (152)	702 (171)	NA
Time after peak velocity (s)	0.72 (0.24)	1.24 (0.49)	G†
max grip aperture (mm)	102 (28)	90 (25)	NA
Time after max grip aperture (s)	0.37 (0.30)	0.91 (0.49)	G†

 Note that the table only reports group effects. G, group effect.

*  Significant at 0.05.

†  Significant at 0.01.

Table A3.

 

Movement Kinematics; Group Mean (±SD) across Visual Group and End-point Location

	Visually Impaired		Normal		ANOVA
	Near End Position	Far End Position	Near End Position	Far End Position
Peak lateral velocity (mm/s)	410 (105)	603 (452)	452 (54)	700 (92)	EP*
Time to peak lateral velocity (s)	0.54 (0.15)	0.64 (0.13)	0.44 (0.08)	0.53 (0.09)	G, EP*
Peak vertical velocity (mm/s)	217 (133)	229 (123)	133 (73)	177 (93)	EP, (g-ep)
Time to peak vertical velocity (s)	0.43 (0.15)	0.46 (0.09)	0.36 (0.08)	0.43 (0.14)	G
Vertical velocity corrections (n)	65 (30)	83 (29)	43 (16)	48 (18)	G, EP*, (g-ep)

 Table shows data averaged across initial start position (IP) and object size (OBJ) to illustrate the effects of vision group (G) and end-position location (EP). Factors found to be significant are shown by capital letter (P < 0.05) and asterisks (P < 0.001), and interactions between factors are denoted by lowercase letter in parentheses.

Table A4.

 

Correlation Coefficients between Movement Kinematics and Visual Function

	VA			CS			Central 5°			Central 10°			Central 20°
	r	r 2	P	r	r 2	P	r	r 2	P	r	r 2	P	r	r 2	P
Peak lateral velocity	−.006	0.00	.975	.263	0.07	.186	.193	0.04	.335	−.299	0.09	.129	−.216	0.05	.279
Time to peak lateral velocity	.392	0.15	.043	−.486	0.24	.010	−.491	0.24	.009	−.590	0.35	.001	−.514	0.26	.006
Peak vertical velocity	.426	0.18	.027	−383	0.15	.049	−.211	0.04	.290	−.129	0.02	.523	−.214	0.05	.284
Time to peak vertical velocity	.601	0.36	.001	−.553	0.31	.003	−.425	0.18	.027	−.476	0.23	.012	−.456	0.21	.017
Vertical deceleration time	.451	0.20	.018	−.690	0.48	.000	−.498	0.25	.008	−.429	0.18	.026	−.406	0.16	.036
Peak vertical height of the wrist	.591	0.35	.001	−.544	0.30	.003	−.375	0.14	.050	−.171	0.03	.394	−.225	0.05	.259

 Significant correlations are denoted in bold.