Clinical Neurophysiology 124 (2013) 19041910Contents lists available at SciVerse ScienceDirect

Clinical Neurophysiology

journal homepage: www.elsevier .com/locate /c l inphGrip force control during simple manipulation tasks in non-neuropathicdiabetic individuals1388-2457/$36.00 2013 International Federation of Clinical Neurophysiology. Published by Elsevier Ireland. All rights reserved.http://dx.doi.org/10.1016/j.clinph.2013.04.002

Corresponding author. Tel.: +55 (11) 3385 3103; fax: +55 (11) 3385 3009.E-mail address: defreitaspb@gmail.com (P.B. de Freitas).P.B. de Freitas , K.C.A. LimaMotion Analysis Lab, Graduate Program in Human Movement Sciences, Cruzeiro do Sul University, Rua Galvo Bueno, 868, Liberdade, So Paulo 01506-000, SP, Brazil

a r t i c l e i n f o h i g h l i g h t sArticle history:Accepted 3 April 2013Available online 3 May 2013

Keywords:Diabetes mellitusHand functionGrip strengthNine hole peg testJebsenTaylorLoad forceObject manipulation The hand function of non-neuropathic diabetic individuals was assessed using traditional hand func-tion tests and instrumented handles.

Performance in two traditional hand function tests and maximum grip strength were not affected bydiabetes.

Surprisingly, non-neuropathic diabetic individuals adopted lower safety margin than controls duringa simple object manipulation.

a b s t r a c t

Objective: To assess hand function and grip force (GF) control in non-neuropathic diabetic individualsusing traditional hand function tests and instrumented handles that provide information about theunderlying neural mechanisms controlling simple manipulation tasks.Methods: Twelve diabetic individuals (3160 years-old) without neuropathy and 12 controls performedtraditional functional tests (i.e., nine hole peg test, JebsenTaylor test, and maximum grip strength test)and were tested for GF control in two situations: holding a free moving instrumented handle and isomet-rically pulling fixed handles. Task performance in the tests and safety margin (SM percentage of GFabove the minimum needed to hold the handle) were the main dependent variables assessed.Results: There was no difference between diabetics and controls in any functional test and in SM in iso-metric pulling task. However, diabetics presented around twice lower SM than controls in the free hold-ing task.Conclusions: Diabetics showed no impairment in functional manipulation tasks. However, they presenteda lower SM than healthy controls.Significance: This lower SM suggests that diabetics may present sensory impairment that could put themat risk of losing objects during its manipulation. Also, it suggests that the applied experimental procedureis sensitive to detect mild sensory impairment in diabetics.

2013 International Federation of Clinical Neurophysiology. Published by Elsevier Ireland. All rightsreserved.1. Introduction

Object manipulation could be considered as an essential func-tional motor action, critical for living an independent lifestyle. Asuccessful object manipulation depends on the individuals abilityto exert an adequate magnitude of grip force (GF force compo-nent acting perpendicularly to the object surface) to prevent slip-page caused by external and by self-generating forces actingtangentially (load force LF) at the digitsobject surface interac-tion. Consistent with a simple mechanical model, in order to holdan object, GF has to be at least equal to the ratio of LF and the staticcoefficient of friction (COF) acting upon the digitsobject interac-tion (i.e., GF = LF/COF) (Johansson and Westling, 1984; Westlingand Johansson, 1984). However, during manipulation individualstend to be more conservative by adopting a safety margin (SM),that is, individuals apply slightly higher GF than the minimumneeded to prevent slippage (GFmin). Also, GF is constantly modu-lated with respect to ongoing changes of LF providing a relativelylow and stable surplus of GF above GFmin (Johansson and Birznieks,2004; Johansson and Flanagan, 2008). This behavior has been ob-served in a variety of manipulation tasks, from holding in placeto shaking a handheld object (de Freitas et al., 2009; Flanaganand Wing, 1995; Zatsiorsky et al., 2005).

http://crossmark.dyndns.org/dialog/?doi=10.1016/j.clinph.2013.04.002&domain=pdfhttp://dx.doi.org/10.1016/j.clinph.2013.04.002mailto:defreitaspb@gmail.comhttp://dx.doi.org/10.1016/j.clinph.2013.04.002http://www.sciencedirect.com/science/journal/http://

P.B. de Freitas, K.C.A. Lima / Clinical Neurophysiology 124 (2013) 19041910 1905It has been generally accepted that the skin mechanoreceptorsprovide information about objects weight and COF and allow forrapid and accurate estimation of the GFmin. Actually, this informa-tion is utilized for a quick adaptation of GF to the current objectsphysical properties and for updating the central controller aboutthe events occurring at the digitsobject surface interaction(Johansson and Birznieks, 2004; Johansson and Flanagan, 2008).It is already known that several neurological diseases alter the cen-tral nervous systems (CNS) ability to control and scale GF with re-spect to LF and COF. For example, mild affected multiple sclerosispatients apply muchmore GF than needed to lift and hold an object(i.e., elevated SM) (Iyengar et al., 2009; Krishnan et al., 2008; Mar-waha et al., 2006). Also, individuals with cerebellar dysfunction(Muller and Dichgans, 1994; Nowak et al., 2002), stroke survivors(Hermsdorfer et al., 2003), individuals with Parkinsons and Hun-tingtons disease (Fellows et al., 1998; Nowak and Hermsdorfer,2002; Serrien et al., 2002), and individual with chronic somatosen-sory deafferentation (Hermsdorfer et al., 2008; Nowak et al., 2004)show an elevation of GF and, consequently, SM when comparedwith healthy individuals in different manipulation tasks. Surpris-ingly, there are no studies about GF control in diabetic individualswithout and with diagnosis of diabetic peripheral sensory neurop-athy (DPN).

According to the World Health Organization (1999) diabetesmellitus (DM) is a metabolic disorder caused by defects in insulinsecretion, insulin action, or both, which directly affect the carbohy-drate, fat and protein metabolism. The DM is characterized bychronic hyperglycemia, which, if persistent, can produce injury,loss of function, and failure of various body tissues and organs.The DM can also cause pathological and functional changes, includ-ing progressive development of retinopathy, nephropathy, and/orneuropathy. Around fifty percent (50%) of diabetic individualsshow some type of neuropathy and the most common is theDPN. The DPN affects the sensory and motor neurons and is char-acterized by the reduction in nerve conduction velocity, decreasedsensitivity in the distal end of upper and lower extremities, and bydecreased motor function in more severe stages (Ramji et al., 2007;Watkins and Thomas, 1998). The DPN remains undetected in mostof the cases and it is diagnosed only by sophisticated clinical andneurological tests (e.g., sensory and/or motor nerve conductionvelocity and electromyography) or when more severe symptomsand complications caused by DPN progress. Symptoms like numb-ness and paresthesias are very common in persons with DPN,mainly in the feet and lower extremities, and they are related tofunctional deficits in the peripheral sensory system. However, de-spite the more severe consequences of the DPN is seen in the lowerextremities (e.g., amputation), the hands are also affected by thedeficits in sensory information (Dahlin et al., 2008).

As most of the diabetic individuals may have subclinical signs ofDPN (e.g., sensory deficits) without presenting any clinical sign andfunctional loss (e.g., maximum power grip strength) (Meijer et al.,2008) and based upon results of previous studies showing thatindividuals with central and peripheral neurological deficits pres-ent changes in GF magnitude control during simple object manip-ulation (Krishnan et al., 2008; Nowak and Hermsdorfer, 2006), webelieve that the GF magnitude could be a sensible performancevariable to detect mild neurological deficits in diabetic individualswithout formal diagnosis of DPN and, consequently, could be usedas the first sign of neuropathy. Therefore, the aim of this study wasto evaluate and compare hand function and GF control of diabeticswithout DPN and healthy controls. We hypothesize that while thetests traditionally used in clinical and research settings to assesshand function would not be sensible to detect differences betweendiabetic individuals without DPN and healthy individuals, the testsusing instrumented handles, which provide accurate informationabout GF control would be able to detect such differences. Specif-ically, we expect that diabetic individuals should select a higherSM than healthy individuals due to slight sensory loss mainly fromthe sensors located at the tip of their digits.2. Methods

2.1. Participants

Twelve diabetic individuals between 31 and 60 years-old(mean SD, 50.3 10.6 years, BMI = 27.53 3.22 kg m2) withoutmedical diagnosis of DPN, and twelve healthy age- and gender-matched controls (49.9 10.55 years-old, BMI = 26.96 3.07 kg m2)volunteered to participate in the study. All participants were right-handed as indicated by their answers to the Edinburg HandednessInventory (Oldfield, 1971). Prior to take part in the study the partici-pants signed an informed consent form approved by the local Institu-tional Research Ethics Committee.

In order to be selected to participate, the diabetic individualsshould not be older than 60 years, be following treatment pre-scribed by a physician, not have diagnosis of DPN, retinopathy,and nephropathy, should not present loss of protective cutaneoussensation in the foot assessed by SemmesWeinstein Monofila-ments Examination (SWME, monofilament 610 g), and shouldhave a score equal or lower than six in the questionnaire part ofthe Michigan Neuropathy Screening Instrument (Feldman et al.,1994; Valk et al., 1994). Both, diabetic individuals and healthy con-trols should be able to understand and follow simple instructionsand have no history of musculoskeletal injury or disease affectingtheir hands (e.g., carpal tunnel syndrome) and upper-extremityfunctions.2.2. Experimental procedure

2.2.1. Hand function assessmentThe experimental procedure started with the examination of

the cutaneous pressure sensitivity of feet (for screening purposes)and hands (i.e., tips of thumb, index and little fingers) using theSWME. After, the participants, comfortably seated in a chair, per-formed three tests traditionally used to evaluate hand function:Rolyan nine hole peg test (9HPT), JebsenTaylor hand function test(JTHFT) and maximum power grip strength (GSMax). The tests wereperformedwith the dominant and non-dominant hands. Half of theindividuals and their respective controls started the tasks withtheir dominant while the other half started with their nondomi-nant hand.

The 9HPT intends to assess digital dexterity and consists ofcatching and placing nine small cylindrical pegs in nine smallholes, one at the time, until all nine holes are filled, followed bythe immediate return of the pegs to their original container(Mathiowetz et al., 1985b). The participants were instructed toperform the task as quick as they could and verbal motivationwas provided during the test execution. They repeated the testthree times with each hand in an alternated way. The time toaccomplish the task was measured by a stopwatch and the shortesttime among the three trials was used as the dependent variable.

The JTHFT is a test designed to evaluate patients hand functionby assessing the performance in tasks (seven subtests) that resem-ble daily executed manipulation tasks (Jebsen et al., 1969). The se-ven subtests are [1] writing short sentences, [2] turning cards, [3]picking and transporting small objects, [4] simulated feeding, [5]stacking checkers, and [6] moving lightweight and [7] heavyweightcans. The first subtest (i.e., writing) was not performed due to thesentence being written in English Idiom and the participants werePortuguese native speakers. The participants were asked to per-form the six subtests as fast as they could and the time of execu-

Fig. 1. (A) Schematic representation of the instrumented handles fixed in theirbases with force transducers depicted as grey cylinders and digits as open ellipsesand (B) a superior view of the participants position and the computer screen duringthe isometric pulling task.

1906 P.B. de Freitas, K.C.A. Lima / Clinical Neurophysiology 124 (2013) 19041910tion of each test was recorded. They performed a single trial foreach subtest as recommended by the test instructions (Jebsenet al., 1969). The dependent variable for this test was the summa-tion of the time spent performing each one.

Lastly, the GSmax was assessed with a Jamar hydraulic handdynamometer (SimmonsPreston Rolyan) according to the normsof the American Society of Hand Therapists (Mathiowetz et al.,1985a). In short, the seated participants were oriented to holdthe dynamometer while keeping their upper arm vertically andthe pronated forearm horizontally oriented, and wrist slightlyhyperextended. Then, they were asked to hold the dynamometerand, after the experimenter go sign, squeeze it as hard as possiblefor 5 s. The force value shown in the gauge of the dynamometer (inkgf) was recorded and the highest value among three trials wasused for statistical analysis.

2.2.2. Grip force control assessmentFor grip force control assessment the participants performed

two different manipulation tasks: isometric pulling task and freeholding task. The tasks were performed with instrumented handles(Fig. 1A). Each handle is composed of two parallel aluminum plates(15 4 1 cm) attached with each other by a compression-ten-sion load cell (LPM-530, Cooper Instruments and Systems, USA)and two aluminum pieces with a multi-axis force and torque (F/T) transducer (Mini40, ATI, USA) in between them, forming the ba-sis of it. The handles were attached on a T shape metal part forthe isometric pulling task. The handle aperture was of 5 cm andits external surface covered with extra fine sandpaper (320 grit).The load cell provides information about the compression force(FC) produced by the tip of the thumb against the handle surface,while the F/T transducer records all three force and torque compo-nents applied against the handle by the digits.

During the isometric pulling task the participants stayed seatedin a height-adjustable chair with one or both shoulders flexed 30,elbows flexed at 120, distal radioulnar joints pronated at 90,wrist slightly hyperextended, and digits slightly flexed. Then, par-ticipants were asked to grasp one or both fixed handles with thetips of the digits (four fingers and the thumb opposing them) eitherwith right, or left, or both hands. The vertically oriented handlewas rotated 45 with respect to the participants frontal plane toassure a comfortable wrist position. Next, they were instructedto isometrically pull the fixed handle(s) up, producing verticalforce (FZ), and keep this force constant to match (i.e., superpose)a horizontal red line set at 6.35 N as accurate as possible. The hor-izontal red line and the current real-time FZ exerted by the partic-ipants were displayed in a 19-in. widescreen computer monitorplaced in front of them (Fig. 1B). FZ was shown as a continuous leftto right running black line in a white background. During bimanualtask, participants were instructed to exert similar upward isomet-ric force with both hands and the FZ exerted by them were aver-aged and shown as a single black line in the monitor. Each triallasted 10 s. Participants performed a first block of three trials tofamiliarize with the task, which ensured adequate task perfor-mance, followed by three trials that were recorded and analyzed.The order of conditions (i.e., unimanual right, unimanual left, andbimanual) was balanced in the diabetic individuals and controls.They only received instructions to exert force upward and GFwas not mentioned during instructions.

In the free holding task the seated participants were firstly askedto grasp, lift and, thereafter, hold the free and vertically orientedhandle weighting 6.35 N. Immediately after the participants foundthemselves in a comfortable position and the handle was stationarythe trial started and data were collected. The participants were in-structed to hold the handle as stable as possible for 10 s as theywould hold a glass of water and after hearing a beep sound theywere instructed to slowly reduce the magnitude of GF until thehandle slips off of the hand. This last procedure (slip test) was donefor identification of GFmin and calculation of the slip ratio (Savescuet al., 2008; Westling and Johansson, 1984). The participants re-peated this procedure five times and the average slip ratio fromthe last three trials was used for further analysis. A high reliabilityof the slip ratio test has already been documented (de Freitas andJaric, 2009; Uygur et al., 2010b). The participants performed thistask only with their dominant hand (right) since we have evidencefrom previous study (de Freitas and Jaric, 2009) and we observed ina pilot study with diabetics and controls that there is no differencebetween hands neither in SM nor in the slip ratio.

2.3. Data processing and analyses

Two customized LabView (National Instruments, USA) routineswere used for data acquisition and processing. The force signalswere recorded at 200 Hz and stored for later analyses. The rawforce signals were low-pass filtered with a 4th-order, zero lag, But-terworth filter with a cut-off frequency set at 20 Hz. Next, GF andLF were calculated. For isometric pulling task, LF exerted againstthe handle was calculated as the resultant force of the two tangen-tial force components [i.e., vertical FZ and horizontal FX,LF =

p(FZ2 + FX2)] and GF was calculated by averaging the force ex-

erted against two sides of the handle (i.e., FC recorded by the loadcell and FY recorded by the F/T transducer) using the equation:GF={[FC + (FC FY)]/2} (de Freitas and Jaric, 2009; Uygur et al.,2010a). In the free holding task, LF was simply the weight of thehandle (i.e., 6.35 N) because the handle was kept stationary andthe horizontal component acting tangentially was negligible.

Table 1Values represent the group means (SD) of the performance in the nine hole peg test(9HPT, in seconds), in the JebsenTaylor hand function test (JTHFT, in seconds), andthe maximum grip strength (GSmax, in kgf) when using dominant and nondominanthand.

Group Hand 9HPT (s) JTHFT (s) GSmax (kgf)

Diabetics Dominant 16.23 26.99 40.2(2.09) (2.96) (5.54)

Nondominant 17.96 29.30 38.6(1.61) (2.50) (6.53)

Controls Dominant 16.07 25.49 45.2(1.77) (0.98) (13.50)

Nondominant 17.12 27.73 41.5(1.79) (2.03) (11.76)

P.B. de Freitas, K.C.A. Lima / Clinical Neurophysiology 124 (2013) 19041910 1907Moreover, GF exerted against the object surface was provided onlyby FC, as normal forces applied on both sides of the handle shouldbe similar in order to stabilize the handle (a task constraint) andmatch the equilibrium requirement.

In the isometric pulling task, although the task lasted 10 s, onlythe interval between the 3rd and the 9th second was analyzed. Thefirst 3 s were considered as a time needed for initial adjustment ofFZ to the prescribed force level, while the final second could be af-fected by expectation of the trial ending. To evaluate task perfor-mance during this task we calculated the root mean square error(RMSE) of the exerted FZ with respect to the target. GF stabilitywas assessed by the coefficient of variation (GF CV) shown in per-centage of averaged GF (GFmean). In addition, GF control was as-sessed by relative SM [SMrel = 100 ((GFmean GFmin)/GFmin)],where GFmean is the averaged GF during the selected time interval(Danion, 2008; de Freitas et al., 2009). Note that GFmin was ob-tained in the free holding task. Despite that, no difference in COFand, consequently, in GFmin is expected during both situations(Savescu et al., 2008).

Regarding the free holding task, we divided it in two phases:holding and slippage phase. The holding phase lasted 10 s, butwe analyzed the central 6 s, skipping the first two and last 2 s.For holding phase we computed CV of GF as well as SMrel adoptingthe same calculation procedures shown above. For slippage phasewe estimated GFmin and calculated the slip ratio (GFmin/LF). TheFig. 2. (A) Index of task performance (RMSE, in N), (B) grip force stability (CV of GF, inaveraged across participants during the isometric pulling task. Error bars represent stanpoint of slippage was determined as the point in time in which asudden reduction of the FZ recorded from the F/T transducer dueto the handles acceleration occurred (Savescu et al., 2008; Uyguret al., 2010b). GFmin (and slip ratio) was estimated from the pointimmediately before the beginning of slippage.

2.4. Statistical analyses

The dependent variables were tested for normal distributionwith ShapiroWilk test. After assuring normal distribution of allvariables, we completed several analyses of variance (ANOVAs).% of the mean GF exerted), and (C) relative safety margin (SMrel, in % of the GFmin)dard deviations.

1908 P.B. de Freitas, K.C.A. Lima / Clinical Neurophysiology 124 (2013) 19041910For hand function assessment, we performed three two-way ANO-VAs (group and hand), with the last factor treated as repeated mea-sure, to test for differences between groups (diabetic individualsvs. healthy controls) and hand (dominant vs. non-dominant) inthe performance of 9HPT and JTHFT, and in the GSmax. For isometricpulling task we performed a single two-way ANOVA (group andtask), with the last factor treated as repeated measure, to test fordifferences between groups and tasks performed (unimanual rightvs. unimanual left vs. bimanual) in the RMSE. Also, two three-wayANOVAs, with the two last factors treated as repeated measure,was carried out to test for differences between group, hand, andtask (unimanual vs. bimanual) in CV of GF and SMrel. Finally, forfree holding task three one-way ANOVAs were performed to testfor differences between groups in slip ratio, CV of GF and SMrel.The level of significance was set at p < .05.

3. Results

3.1. Cutaneous pressure sensitivity in diabetics

The cutaneous pressure sensitivity in the tip of the thumb, in-dex and minimum fingers was assessed with SWME. Accordingto the results of this examination, all 12 control participants and9 out of 12 diabetic individuals showed no observable reductionin cutaneous pressure sensitivity in their dominant and non-dom-inant hands, being able to feel the tiniest monofilament (0.05 g).Only three diabetic individuals showed small decrease of cutane-ous sensation, not being able to feel the tiniest monofilament inall tested digits, but being capable of feeling the next one (0.2 g).

3.2. Traditional hand function tests

Traditional hand function tests were applied and results re-vealed no difference between diabetic individuals and healthy con-trols for any test [9HPT: F(1,22) = .51, p > .05, g2 = .02; JTHFT:F(1,22) = 2.41, p > .05, g2 = .1; and GSMax: F(1,22) = .68, p > .05,g2 = .03]. Also, no group by hand interaction was revealed(p > .05). Results only revealed an effect of hand for all three tests,where participants performed better with their dominant hands[9HPT: F(1,22) = 32.4, p < .001, g2 = .60; JTHFT: F(1,22) = 35.1,p < .001, g2 = .62; and GSMax: F(1,22) = 7.5, p < .05, g2 = .25]. Table 1depicts mean and standard deviation of the performance in 9HPTand JTHFT, as well as GSMax.

3.3. Isometric pulling task

During isometric pulling, the task performance was assessed byRMSE (Fig. 2A). The results revealed no difference between groupsfor RMSE [F(1,22) = .04, p > .05, g2 = .01]. Moreover, there were nei-Fig. 3. (A) Slip ratio, (B) relative safety margin (SMrel, in % of the GFmin), and (C) the coewhen participants held the free moving handle with the dominant hand. Error bars repther effect of task nor group by task interaction (p > .05). RegardingGF steadiness (Fig. 2B), no difference between groups was foundfor CV of GF [F(1,22) = 1.15, p > .05, g2 = .05]. Also, no group byhand, group by task, and group by hand by task interactions wererevealed (p > .05). Similarly, no main effect of hand and of task andno hand by task interaction were revealed (p > .05). About GF con-trol (Fig. 2C), the results revealed that diabetic individuals pre-sented similar SMrel than controls [F(1,22) = 2.95, p = .1, g2 = .12].Moreover, no group by hand, and group by hand by task interac-tions were revealed for SMrel (p > .05). Also, no main effect of handand of task and no interaction between them were revealed(p > .05).

3.4. Free holding task

The free holding task was divided in two phases, holding andslippage phase. During slippage phase the only variable of interestwas the slip ratio that did not differ between diabetics and controls[F(1,22) = 0.19, p > .05, g2 = .01]. Regarding GF steadiness, resultsrevealed no difference between groups [F(1,22) = .73, p > .05,g2 = .03] for CV of GF. However, results indicated that SMrel wastwice smaller for diabetic individuals when compared to controls[F(1,22) = 13.18, p < .005, g2 = .38]. Fig. 3 depicts means and respec-tive standard deviations of slip ratio, CV of GF and SMrel.

It was mentioned above that three diabetic individuals pre-sented slight reduction in cutaneous pressure sensitivity assessedby SWME. Therefore, we ranked them according to their SMrel inorder to examine whether the reduction in sensitivity would affectthe GF magnitude exerted while holding a free moving object. Twodiabetic individuals with reduced cutaneous sensitivity presentedthe second and the third lowest SMrel, whereas one diabetic withreduced sensitivity presented the second largest SMrel, whichwas smaller but close to the mean SMrel presented by the controlgroup (126.4% and 133.5%, respectively).

4. Discussion

The aim of this study was to assess different aspects (i.e., clinicaltests performance and underlying neural control mechanisms) in-volved in hand function in diabetic individuals without peripheralneuropathy and healthy controls. The results indicated that thediabetic individuals had no major loss of cutaneous pressure sensi-tivity in their hands as revealed by the SWME. Also, results indi-cated that the diabetic individuals had similar performance inmanipulation tasks involving digits dexterity (9HPT) and actionsof the whole hand and upper extremity (JTHFT) when comparedwith controls. Likewise, both groups generate similar maximumpalmar grip strength. Concerning the control of hand function,we found the same trend of similarities between diabeticfficient of variation of the exerted grip force (CV of GF, in % of the mean GF exerted)resent standard deviations.

P.B. de Freitas, K.C.A. Lima / Clinical Neurophysiology 124 (2013) 19041910 1909individuals and controls. Indeed, we found that diabeticindividuals had similar task performance than controls when askedto exert a constant amount of force by pulling an instrumentedhandle up and superposing a horizontal line presented in a com-puter monitor. It was also found that diabetic individuals kept GFrelatively stable (i.e., low CV of GF) when performing two simplemanipulation tasks with constant tangential force, being compara-ble with controls. However, the groups were different in a single,but important, dependent variable. Conflicting with the hypothe-sis, we found that diabetic individuals presented lower safetymargin (SM) relative to the GFmin, when performing a simplemanipulation task (i.e., free holding task), when compared tocontrols and presented a trend of lower SM when performing anisometric pulling task. This intriguing low SM set by diabeticindividuals deserves special attention and will be the focus of thediscussion hereafter.

It is already known that afferent signals coming from cutane-ous mechanoreceptors located at the glabrous skin of the tip ofthe digits are crucial for providing the central controller informa-tion about the current state (e.g., magnitude of GF and occurrenceof microslips) at the digitsobject interaction and for updating thecontroller about the necessary GF magnitude for keeping a safeand stable grasp (Flanagan and Wing, 1995; Johansson andWestling, 1984). In general, when the tip of the digits are anes-thetized and, consequently, there is partial or complete attenua-tion of sensory inputs going from the periphery to the center,the central controller sends motor commands to the peripheryincreasing GF magnitude during a simple lifting task (Augurelleet al., 2003a,b; Johansson and Westling, 1984; Monzee et al.,2003). However, when individuals are asked to hold the objectfor a relatively long time (i.e., 20 s) SM reduces at the end ofthe holding phase as compared to the beginning (Augurelleet al., 2003b). This reduction happens without and with anesthe-sia, which indicates that the central controller adjusts the GFmagnitude seeking a more economical solution for GF exertion.Nevertheless, the reduction in GF is steeper after anesthesia,which signifies that the lack of reliable sensory information com-ing from the tip of the digits impairs the ability of the centralcontroller to properly regulate the GF magnitude, increasing therisk of object slippage in a longer run (Augurelle et al., 2003b).In the present study, we did not selected specific points in timeto obtain GF and calculate SM as done by Augurelle et al.(2003b). Instead, we calculated the averaged GF during the mid-dle section of the trial. Therefore, we could not evaluate if therewas a similar trend for reduction in SM over time. However, wecalculate the CV of GF to assess the changes in GF during eachtrial and we could not find any difference between diabeticsand controls in GF variability. Thus, we could assume that if therewas a reduction in SM over time, this reduction would be similarin diabetic individuals and controls and could not explain thelower SM in diabetic individuals as compared to controls. Hence,based upon results of studies that used anesthesia to reduce theinflow of sensory information, we could not find a plausibleexplanation about this intriguing and even counterintuitivereduction in SM in diabetic patients, as anesthetized individuals,despite reducing SM over time, still presented higher SM thanwhen they were not anesthetized (Augurelle et al., 2003b).

In healthy individuals with preserved musculoskeletal andcentral and peripheral nervous systems, an adequate GF controlis achieved by a complex interaction between feedforward andfeedback control mechanisms. The amount of GF is set in advancebased on previous experience with the manipulated object andduring the manipulation GF magnitude is continuously adjustedbased on sensory information coming mainly from the tip ofthe digits in contact with the object surface (Flanagan and Wing,1995; Johansson and Westling, 1984). However, in individualswith neurological diseases affecting their CNS changes in GF con-trol have been observed. For example, individuals with multiplesclerosis (Iyengar et al., 2009; Krishnan et al., 2008; Marwahaet al., 2006), Parkinsons disease (Fellows et al., 1998; Nowakand Hermsdorfer, 2002), and stroke (Hermsdorfer et al., 2003)apply much more GF than healthy individuals in a number ofmanipulation tasks. Despite having different origins, the CNS ofindividuals with neurological diseases sets the same solution,i.e., it increases SM while those individuals are manipulating ob-jects. In neurological individuals, the increased SM could be dueto permanent central and, in some cases, peripheral neurologicaldamage. As the control system should be able to detect structuraland functional changes in its components, the system of thoseindividuals would adopt a conservative and compensatory GFcontrol strategy increasing GF in order to prevent slippage thatcould be caused by unexpected LF changes.

However, when the peripheral nervous system is affectedchanges in GF control may or may not happen. Some studies haveshown that individuals with severe carpal tunnel syndrome (CTS),who have cutaneous sensitivity impairment, exert larger magni-tude of GF than controls (Lowe and Freivalds, 1999; Zhanget al., 2013) during manipulation of objects of different masses.Conversely, some studies have shown that individuals with CTS(Thonnard et al., 1999) or individuals who have small reductionin cutaneous sensation caused by a mild compression of the med-ian nerve, which mimics mild CTS (Cole et al, 2003), present nochange in GF magnitude control during simple manipulationtasks. In addition, two studies from Nowaks group (Nowaket al., 2003; Nowak and Hermsdorfer, 2003) using the same groupof individuals with moderately impaired cutaneous sensitivityshowed conflicting results regarding GF scaling during manipula-tion tasks. While Nowak et al. (2003) found that individuals withimpaired sensitivity exerted higher GF than controls when liftingand holding an instrumented object, Nowak and Hermsdorfer(2003) found that this difference between groups was not pre-sented during a point-to-point task. Despite the controversy,someone, based upon the results of the studies that assessed GFcontrol in individuals with peripheral sensory impairments,would expect that diabetic individuals would either present high-er or similar SM when compared to healthy controls. Nonetheless,the present study was the first one to show that a group of indi-vidual with a diagnosed disease that could affect the nervous sys-tem produces less GF than a group of age and sex-matchedcontrol individuals. Then, what would be the reason for diabeticindividuals without neuropathy employ a low SM while holdinga free moving object?

We consider that this low SM in diabetic individuals would be asign of very mild deficit in cutaneous sensitivity that would not beidentified by SWME, which is a conscious discriminatory task andlimited in terms of resolution, but would be important and detri-mental for GF control during manipulation of objects. This lowSM in diabetic individuals could also be a sign of a very mild andundetected deficits in upper limb proprioception (e.g., muscle spin-dle), which is known to be important in GF control (Danion, 2007).Hence, we suggest that this mild sensory loss faced by diabeticindividuals is not sufficient to trigger the use of compensatory GFcontrol strategies as it is in individuals with diseases affectingthe CNS and in individuals with moderate and more severe lossof cutaneous sensitivity, but this very mild sensory loss is sufficientto disrupt the processing and proper use of sensory information inthis population causing an error in the estimation of GF neededduring object manipulation. Nevertheless, we are aware that thissuggestion is speculative and has no support in other studies dueto the novelty of the findings. Therefore, we believe that morestudies are needed to confirm this proposition or to provide alter-native explanation for this phenomenon.

1910 P.B. de Freitas, K.C.A. Lima / Clinical Neurophysiology 124 (2013) 190419105. Conclusion

In conclusion, diabetic individuals without peripheral neuropa-thy show no worsening in hand and upper-extremity function andare able to produce as much maximum grip strength as controls.They also have similar task performance than controls when theyare asked to superpose a target exerting upward isometric verticalforce as well as they are able to keep GF stable throughout thetasks. However, diabetic individuals exert less GF than controlswhen performing simple manipulation tasks, keeping a low SM.This low SM could indicate that diabetic individuals have very mildsensory deficits that are not large enough to make the CNS triggersa compensatory control strategy that would increase GF and SMbut enough to cause error in the estimation of GF needed duringobject manipulation. The low SM could be detrimental for diabeticindividuals while manipulating objects and would put them at riskof losing a handheld object. Finally, the results suggest that thiskind of evaluation, in which a person needs only to grasp and holdan instrumented handle and has GF recorded, could be much moresensible to identify mild sensory deficits than clinical tests used forcutaneous sensitivity assessment (i.e., SWME). Nonetheless, a largescale study needs to be performed in order to assess the feasibilityof this procedure to detect mild and more severe changes in sen-sory and, also, motor systems in individuals with diabetes mellitusand those who are affected by peripheral neuropathy.

Acknowledgments

The authors are thankful to the Sao Paulo State Research Foun-dation (FAPESP, Sao Paulo, Brazil) for its financial support for thisresearch (Grant FAPESP #2010/02939-4). K.C.A. Lima is thankfulfor his scholarship provided by Coordination for the Improvementof Higher Education Personnel (CAPES Brazil).

References

Augurelle AS, Penta M, White O, Thonnard JL. The effects of a change in gravity onthe dynamics of prehension. Exp Brain Res 2003a;148:53340.

Augurelle AS, Smith AM, Lejeune T, Thonnard JL. Importance of cutaneous feedbackin maintaining a secure grip during manipulation of hand-held objects. JNeurophysiol 2003b;89:66571.

Cole KJ, Steyers CM, Graybill EK. The effects of graded compression of the mediannerve in the carpal canal on grip force. Exp Brain Res 2003;148:1507.

Dahlin LB, Thrainsdottir S, Cederlund R, Thomsen NOB, Eriksson KF, Rosen I, et al.Vibrotactile sense in median and ulnar nerve innervated fingers of men withType 2 diabetes, normal or impaired glucose tolerance. Diabet Med2008;25:5439.

Danion F. The contribution of non-digital afferent signals to grip force adjustmentsevoked by brisk unloading of the arm or the held object. Clin Neurophysiol2007;118:14654.

Danion F. Grip force safety margin in rock climbers. Int J Sports Med 2008;29:16872.

de Freitas PB, Jaric S. Force coordination in static manipulation tasks performedusing standard and non-standard grasping techniques. Exp Brain Res2009;194:60518.

de Freitas PB, Uygur M, Jaric S. Grip force adaptation in manipulation activitiesperformed under different coating and grasping conditions. Neurosci Lett2009;457:1620.

Feldman EL, Stevens MJ, Thomas PK, Brown MB, Canal N, Greene DA. A practicaltwo-step quantitative clinical and electrophysiological assessment for thediagnosis and staging of diabetic neuropathy. Diabetes Care 1994;17:12819.

Fellows SJ, Noth J, Schwarz M. Precision grip and Parkinsons disease. Brain1998;121(9):177184.

Flanagan JR, Wing AM. The stability of precision grip forces during cyclic armmovements with a hand-held load. Exp Brain Res 1995;105:45564.

Hermsdorfer J, Hagl E, Nowak DA, Marquardt C. Grip force control during objectmanipulation in cerebral stroke. Clin Neurophysiol 2003;114:91529.

Hermsdorfer J, Elias Z, Cole JD, Quaney BM, Nowak DA. Preserved and impairedaspects of feed-forward grip force control after chronic somatosensorydeafferentation. Neurorehabil Neural Repair 2008;22:37484.Iyengar V, Santos MJ, Ko M, Aruin AS. Grip force control in individuals with multiplesclerosis. Neurorehabil Neural Repair 2009;23:85561.

Jebsen RH, Taylor N, Trieschmann RB, Trotter MJ, Howard LA. An objective andstandardized test of hand function. Arch Phys Med Rehabil 1969;50:3119.

Johansson RS, Birznieks I. First spikes in ensembles of human tactile afferents codecomplex spatial fingertip events. Nat Neurosci 2004;7:1707.

Johansson RS, Flanagan JR. Tactile sensory control of object manipulation inhumans. In: Basbaum AI, Kaneko A, Shepherd GM, Westheimer G, editors.The senses: a comprehensive reference. San Diego: Academic Press; 2008. p.6786.

Johansson RS, Westling G. Roles of glabrous skin receptors and sensorimotormemory in automatic control of precision grip when lifting rougher or moreslippery objects. Exp Brain Res 1984;56:55064.

Krishnan V, de Freitas PB, Jaric S. Impaired object manipulation in mildly involvedindividuals with multiple sclerosis. Motor Control 2008;12:320.

Lowe BD, Freivalds A. Effect of carpal tunnel syndrome on grip force coordination onhand tools. Ergonomics 1999;42:55064.

Marwaha R, Hall SJ, Knight CA, Jaric S. Load and grip force coordination in staticbimanual manipulation tasks in multiple sclerosis. Motor Control 2006;10:16077.

Mathiowetz V, Kashman N, Volland G, Weber K, Dowe M, Rogers S. Grip andpinch strength: normative data for adults. Arch Phys Med Rehabil1985a;66:6974.

Mathiowetz V, Weber K, Kashman N, Volland G. Adult norms for the nine hole pegtest of finger dexterity. OTJR 1985b;5:2438.

Meijer JWG, Lange F, Link TP, van der Hoeven JH. Muscle fiber conductionabnormalities in early diabetic polyneuropathy. Clin Neurophysiol2008;119:137984.

Monzee J, Lamarre Y, Smith AM. The effects of digital anesthesia on force controlusing a precision grip. J Neurophysiol 2003;89:67283.

Muller F, Dichgans J. Dyscoordination of pinch and lift forces during grasp inpatients with cerebellar lesions. Exp Brain Res 1994;101:48592.

Nowak DA, Hermsdorfer J. Coordination of grip and load forces during verticalpoint-to-point movements with a grasped object in Parkinsons disease. BehavNeurosci 2002;116:83750.

Nowak DA, Hermsdorfer J. Selective deficits of grip force control during objectmanipulation in patients with reduced sensibility of the grasping digits.Neurosci Res 2003;47:6572.

Nowak DA, Hermsdorfer J. Objective evaluation of manual performance deficits inneurological movement disorders. Brain Res Rev 2006;51:10824.

Nowak DA, Hermsdorfer J, Marquardt C, Fuchs HH. Grip and load force couplingduring discrete vertical arm movements with a grasped object in cerebellaratrophy. Exp Brain Res 2002;145:2839.

Nowak DA, Hermsdorfer J, Marquardt C, Topka H. Moving objects with clumsyfingers: how predictive is grip force control in patients with impaired manualsensibility? Clin Neurophysiol 2003;114:47287.

Nowak DA, Glasauer S, Hermsdorfer J. How predictive is grip force control in thecomplete absence of somatosensory feedback? Brain 2004;127:18292.

Oldfield RC. The assessment and analysis of handedness: the Edinburgh inventory.Neuropsychologia 1971;9:97113.

Ramji N, Toth C, Kennedy J, Zochodne DW. Does diabetes mellitus target motorneurons? Neurobiol Dis 2007;26:30111.

Savescu AV, Latash ML, Zatsiorsky VM. A technique to determine friction at thefingertips. J Appl Biomech 2008;24:4350.

Serrien DJ, Burgunder JM, Wiesendanger M. Control of manipulative forces duringunimanual and bimanual tasks in patients with Huntingtons disease. Exp BrainRes 2002;143:32834.

Thonnard J, Saels P, Van den Bergh P, Lejeune T. Effects of chronic median nervecompression at the wrist on sensation and manual skills. Exp Brain Res1999;128:614.

Uygur M, de Freitas PB, Jaric S. Effects of varying the load force range and frequencyon force coordination in static manipulation. Neurosci Lett 2010a;475:1159.

Uygur M, de Freitas PB, Jaric S. Frictional properties of different hand skin areas andgrasping techniques. Ergonomics 2010b;53:8127.

Valk GD, Grootenhuis PA, Bouter LM, Bertelsmann FW. Complaints of neuropathyrelated to the clinical and neurophysiological assessment of nerve function inpatients with diabetes mellitus. Diabetes Res Clin Pract 1994;26:2934.

Watkins PJ, Thomas PK. Diabetes mellitus and the nervous system. J NeurolNeurosurg Psychiatry 1998;65:62032.

Westling G, Johansson RS. Factors influencing the force control during precisiongrip. Exp Brain Res 1984;53:27784.

World Health Organization. Definition, diagnosis and classification of diabetesmellitus and its complications. Report of a WHO Consultation, Geneva: WHO;1999.

Zatsiorsky VM, Gao F, Latash ML. Motor control goes beyond physics: differentialeffects of gravity and inertia on finger forces during manipulation of hand-heldobjects. Exp Brain Res 2005;162:3008.

Zhang W, Johnston JA, Ross MA, Sanniec K, Gleason EA, Dueck AC, et al. Effects ofcarpal tunnel syndrome on dexterous manipulation are grip type-dependent.PLoS ONE 2013;8:e53751.

Grip force control during simple manipulation tasks in non-neuropathic diabetic individualsIntroductionMethodsParticipantsExperimental procedureHand function assessmentGrip force control assessment

Data processing and analysesStatistical analyses

ResultsCutaneous pressure sensitivity in diabeticsTraditional hand function testsIsometric pulling taskFree holding task

DiscussionConclusionAcknowledgmentsReferences