Abstract We investigated the effects of old age on the fingertip force responses that occurred when a grasped handle was pulled unexpectedly to increase the ...
Exp Brain Res (2001) 136:535–542 DOI 10.1007/s002210000613
R E S E A R C H A RT I C L E
Kelly J. Cole · Diane L. Rotella
Old age affects fingertip forces when restraining an unpredictably loaded object
Received: 10 May 2000 / Accepted: 11 October 2000 / Published online: 16 December 2000 © Springer-Verlag 2000
Abstract We investigated the effects of old age on the fingertip force responses that occurred when a grasped handle was pulled unexpectedly to increase the tangential load at the fingertip. These automatic responses, directed normal to the handle surface, help prevent slips between the handle and finger. Old adults (average age 78 years) responded with large peak fingertip forces compared to young adults (average age 30 years), even though the two subject groups showed similar skin slipperiness. For step-shaped loads the average response latency was the same for young and old subjects (about 80 ms). Thus, these automatic responses are not susceptible to the age-related central delays known for simple reaction-time tasks. For ramp-shaped loads the average response latency was inversely related to load rate. Response latency was 25 ms longer for the Old group versus the Young group for loads of 8 N/s, and this difference increased exponentially to a 110-ms difference for 2-N/s loads. A twofold difference in the tangential force required to evoke a response was predicted from linear regressions and can account for the latency difference (0.2 N vs 0.4 N threshold for young and old, respectively, r=0.93 for both groups). This theoretical elevation in load force threshold is consistent with degraded central information processing in old age, and the deterioration of cutaneous mechanoreceptors. Keywords Old age · Hand · Motor control · Cutaneous · Sensory
Introduction Cutaneous sensory function of the hands deteriorates in old age, as indicated by less efficient mechanoelectric transduction (Schmidt et al. 1990) and decreased tactile sensibility (Kenshalo 1986; Gescheider et al. 1994; K.J. Cole (✉) · D.L. Rotella Department of Exercise Science, 501 Field House, The University of Iowa, Iowa City, IA 52242, USA Tel.: +1-319-3359491, Fax: +1-319-3356966
Stevens et al. 1998). These declining abilities are consistent with reports of reduced populations of Meissner’s and Pacinian corpuscles in glabrous skin, and changes in receptor morphology (Dickens et al. 1963; Cauna 1965; Bolton et al. 1966). The ability to use available tactile signals may be affected further by a general decline in the capacity to process sensory information centrally (Cerella 1990). These age-related declines in sensory function may explain why old adults used exceedingly large grip forces when frictional properties were unpredictable in objects that they lifted, but not when object weight varied unpredictably (Cole et al. 1999). We suspect this because Meissner’s corpuscles and their associated fast-adapting afferents (FA I) alone encode the frictional properties of grasped objects (Johansson and Westling 1987; Westling and Johansson 1987). A degraded capacity to encode and/or process FA I afferent signals predicts that old adults also will be challenged when handling objects that exert unpredictable changes in load on the hand (e.g., operating a sticky drawer, holding a child’s hand). When grasped objects were pulled upon unpredictably in healthy young adults (increasing the load tangential to the fingertip), the force normal to the object surface increased with a minimum response latency of 70–80 ms (Cole and Abbs 1988; Johansson et al. 1992b). Response latency was inversely related to the load rate, consistent with a load threshold of 0.2 N to evoke a response, and the peak rate of the normal force response scaled in proportion to load rate (Johansson et al. 1992b). There is good evidence that discharges from FA I afferents provide the earliest information for triggering these responses and for scaling response magnitude (Macefield et al. 1996a; Macefield and Johansson 1996). Across cutaneous, muscle and joint afferents, FA I afferents discharged first after load onset, were most rapidly recruited, and encoded tangential load rate independent of grip force magnitude. Also, cutaneous anesthesia of the finger and thumb yielded delayed and frequently absent responses that did not scale well with load rate (Cole and Abbs 1988; Johansson et al. 1992a).
536
We examined the forces produced at the fingertip by young and old adults when pulling loads were applied unexpectedly to a handle that subjects restrained by pressing against it with the fingertip. The size of the force response directed normal to the handle surface, and response latency, were measured across different sizes and rates of pulling loads.
Materials and methods Subjects, medical histories and diagnostic tests Subjects were grouped for data analysis into a “Young” group (six females and three males; ages 21–37 years, mean 29.6 years) and an “Old” group (eight females and seven males; ages 70–88 years, mean 77.8 years). The appropriate internal review board approved participant selection and the experiment was conducted in accord with the Declaration of Helsinki. All subjects reported their medical history (including current medications), occupational history, activities and hobbies. The Young group indicated no health problems. All subjects in the Old group reported themselves as healthy, ambulatory and living independently. Upon arriving at the laboratory for testing, all Old subjects walked unaided, and appeared to be healthy, vigorous individuals. Six Old subjects reported diseases common for their age. In all cases severity was reported to be “mild,” causing little or no restriction of daily activities. They reported hypertension (n=4), heart disease (n=2) and adult-onset diabetes (n=1). No subject reported upper extremity numbness, tingling or pain during the experimental protocol. Two females (ages 86 years) disclosed previous signs and symptoms of carpal tunnel syndrome, but reported freedom from symptoms beginning several years previously. They did not show Phalen’s or Tinel’s signs for median nerve injury. One subject (female, age 85 years) suffered from pernicious anemia several years previously, but reported no sensory dysfunction since her medical treatment. One subject (age 79 years, male) exhibited a positive Tinel’s sign in the third digit of the dominant hand, but no Phalen’s sign in either hand. Five subjects reported “mild” finger joint stiffness, but only one subject reported taking medication (ibuprofen, non-prescription). We used Semmes-Weinstein pressure filaments (Smith & Nephew Roland Inc.; Menominee Falls, WI) to obtain tactile pressure thresholds from the right index finger and thumb, as described previously (Cole et al. 1999). Apparatus A servo-controlled torque motor (rotary moving-coil motor, Series 310B Servo System, Aurora Scientific Inc.; Aurora, Ontario, Canada) was equipped with a handle attached to the motor spindle. This system, running under force-feedback control, delivered pulling loads to the index finger that were directed distally, along the long axis of the finger (Fig. 1A). A laboratory computer and 16-bit digital-to-analog conversion board provided the command signals to the servo system. The system had a force signal resolution of 0.01 N, a linearity of 0.2%, and a force step response of 1.0 ms (1–99%, critically damped). Handle position was transduced via a capacitive detector. Length signal linearity was 0.1%, with a 1-µm resolution for the 8-cm handle length. A single-axis miniature accelerometer (GY-125-5; Kulite Semiconductor Products Inc., Leoria, NJ) was attached near the distal end of the handle, and oriented to transduce tangential handle acceleration. A six-axis force-torque transducer (ATI Nano Force/Torque System; Assurance Technologies Inc.; Garner, NC) was affixed to the distal end of the handle. A flat Plexiglass disk (33 mm diameter) was mounted on the force-torque transducer to serve as the contact surface for the distal volar pad of the index finger. This contact surface was covered with sandpaper (320-grit). We record-
Fig. 1 A Diagram showing the apparatus, direction of tangential load, and grip configuration. B Examples of “ramp” (2, 4, 8 N/s and amplitudes of 2, 4, 6 N) and step-shaped (110 N/s) pulling loads experienced at the fingertip (single trials from one subject)
ed and analyzed the fingertip force that was normal to the contact surface, and the force tangential to the contact surface (horizontal component). The transducer provided a resolution of 0.025 N for the tangential force, and 0.05 N for the normal force. Procedures The subjects sat comfortably beside the apparatus. Their right arm was abducted approximately 30° at the shoulder with the elbow flexed at about 90°. The forearm rested on a rigid support in neutral, extended position. Two canvas tubes filled with lead shot were placed on either side of the forearm to stabilize the limb. Subjects placed their thumb pad against a flat rigid support and restrained the handle by pressing the distal volar pad of their index finger against the sandpaper-covered handle surface (Fig. 1A). We ensured that subjects did not hook their index finger over the distal edge of the handle. The distance between the pads of the index finger and thumb was approximately 5 cm. With their ulnar digits subjects gripped a rigid steel cylinder (5 cm height, 2.5 cm diameter). The responses to tangential loads applied to a single digit (Cole and Johansson 1993; Häger-Ross et al. 1996; Burstedt et al. 1997; Ohki and Johansson 1999) are virtually identical to responses obtained from simultaneous loading of the thumb and index finger (Cole and Abbs 1988; Johansson et al. 1992b, 1992c; Macefield et al. 1996a; Macefield and Johansson 1996). We instructed subjects to press gently against the handle with their fingertip and minimize handle movement during the pulling loads. We did not tell them how this should be accomplished, except to “relax their finger pressure” if their normal force exceeded a few newtons between trials. Each trial began with no tangential load at the handle for several seconds followed by an unannounced increase in pulling load (Figs. 1B, 2). The target load (see below) was maintained for 3–5 s after which the load decreased to 0 N. The time between trials typically was less than 10 s. Subjects did not view their hand while a trial was in progress. The experiment began with five practice trials using loads that increased at 4 N/s to a target load of 2 N. The next 60 trials were randomly ordered across the following load stimuli (ten each): 2 N @ 2 N/s; 2 N @ 4 N/s; 2 N @ 6 N/s; 4 N @ 2 N/s; 4 N @ 4 N/s; and 6 N @ 4 N/s loads (Fig. 1B). Ten subjects (three females and two males: 29–37 years; and two females and three males: 70–79 years) performed two additional blocks of trials employing 2-N loads @ 8 N/s (n=8) and 2 N @ 110 N/s (n=5). The
537 Fig. 2 Examples of signals obtained during a single trial from a young subject. Interval a indicates the latency from tangential load onset to the start of the normal force response. Marker b indicates the peak normal force rate during the rapid phase of the response (“catch-up”). Finger stiffness was calculated from load force and handle displacement changes between the end of the plateau (c) and the beginning of normal force reduction (d)
110-N/s load rate resulted from a step increase in voltage to the servo controller (Fig. 1B); it was chosen to ensure that responses of minimum latency were evoked. Minimum response latency (70–80 ms in young, healthy adults) can be achieved with a tangential load rate of 32 N/s (Johansson et al. 1992b). We obtained estimates of the coefficient of static friction at the distal pad of the index finger (Cole and Johansson 1993). The handle was servo controlled using position feedback to a vertical alignment. We instructed each subject to apply a 2-N pressing force with their index finger against the grip surface, as indicated by visual feedback of normal force on an oscilloscope. While maintaining this force, the subject slowly dragged their index finger across the grip surface horizontally and in a proximal direction. Subjects typically produced several slips within a single trial. The ratio of the tangential force to normal force at slip, averaged across all slips, provided an estimate of the coefficient of static friction for each subject.
We first separated the data into old adults with no relevant disease history (N=9) and those who reported current or previous relevant diseases (as noted previously; N=6). These two groups were compared on all measures of response latency, normal force and normal force rate. For these groups we also compared the parameters of linear regression equations that describe the relationship between response latency and the inverse of the load force rate for each subject (that is, minimum response latency and the theoretical load force threshold; see “Results” and Fig. 4C). There were no statistically significant differences between the groups of old subjects on any measure. Both groups of old subjects showed identical patterns of results when compared separately against the Young group. Therefore, the old subjects were combined into a single group for all comparisons with young subjects that are reported in “Results.”
Results Data collection and analysis Data collection and analysis were accomplished with a laboratory computer system (SC/ZOOM; Department of Physiology, University of Umeå, Sweden) and a personal computer. All signals were collected with 12-bit resolution at 400 samples/s. Normal force rate was obtained using symmetrical numerical time differentiation (±5 point). We measured the latency of the normal force response that followed the load increase (Fig. 2, interval “a”). This latency was the time between the beginning of the rise in horizontal tangential force and the time at which the normal force rate rose above the peak rate observed during the first 50 ms of the load ramp. This criterion did not bias the latency measures for either group because old and young adults were equally stable in their application of fingertip force to the handle during the first 50 ms of the load ramp (also see Cole and Beck 1994). The peak normal force rate (Fig. 2, “b”) provided information on response scaling. Finally, the net “stiffness” of the index finger at the end of the load plateau was estimated by measuring the change in load force and handle position between the decrease in the force ramp and the subsequent decrease in normal force (Fig. 2, “c” and “d”). Data reported herein reflect within-subject mean values averaged across subjects. Analyses of variance (ANOVA) and covariance (ANCOVA), with repeated measures within subject, were used to assess statistical significance. All statistical tests used Statistica for Windows (version 5, StatSoft Inc., Tulsa, OK, USA, 1996) and required a probability level of 0.05 for statistical significance.
Both subject groups began each trial with a small normal force against the handle (0.62 and 0.78 N normal force for Young and Old, respectively). They increased their normal force after the tangential loading began (Figs. 2, 3), exhibiting the characteristics described by Johansson et al. (1992b, 1992c). Specifically, the response latency (Fig. 2, interval “a”) decreased as load rate increased, the rapid initial phase of the normal force response lasted about 300 ms (Fig. 2, “catch-up”), and the peak normal force rate (Fig. 2, point “b”) scaled in proportion to load rate. On trials when the tangential force continued to increase after the rapid phase of the grip response was completed, subjects increased their normal force steadily, paralleling the rising tangential force (Fig. 2, “track”). Subjects in the Old group showed longer response latencies than the Young group for all loads except the step-shaped loads (110 N/s; Fig. 4). The size of this difference depended upon the load rate. For the fastest loads (110 N/s) subjects in the two groups exhibited similar response latencies (76 vs 80 ms for Young and Old, respectively; P>0.26). Beginning with the 8-N/s loads, a significant latency difference between Young and Old was observed (110 ms vs 135 ms, P
