Cognitice Brain Research, 1 f 1993) 197-201 0 1993 Elsevier Science Publishers B.V. All rights reserved 092W~410/93/$06.80

197

BRESC 4tW4

Intensity to force translation: a new effect of stimulus-response compatibility revealed by analysis of response time and electromyographic activity of a prime mover Patricia RomaiguGre a, Thierry Hasbroucq b, Camille-Aim6 Possamdi ’ and John Seal b u Equi@ede Physioogie NeuromusculaireHumaine, LaboraroireNeumbiologie et Moucements, Centre Na&nai de la RechercheS&n@que, ’ Laboratoire de NeurosciencesCognitices, Centre National de la Recherche Scientifique, c Labomtoire de NeurosciencesFoncknmeUes. Centre National de la Recherche Scientifique, Marseille (France) (Accepted 9 March 1993)

Key words: Fractionated reaction time; Visual stimulus intensity: Isometric contraction; Man

In reaction time studies of stimulus-response corn~+ibttt~ r ______._~,emphasis bar been placed on the influence of spatial stimulus-response relationships. but what seems to be essential for the emergence of an effect of stimulus-response compatibility is the existence of a conceptual match between stimulus and response variables. This notion was at the origin of the present study to assess the compatibility relationship between the intensity of a visual stimulus and the force of a voluntary muscle contraction. A stimulus-response compatrbility effect was demonstrated. This effect was entirely due to premotoric processes.

One of the most reliable results of informationprocessing psychology is that the relationship between stimulus 6) and response CR) variables is a critical %terminant of human performance4. A classical manipulation of the S-R relationship consists of varying, by way of different task instructions, the mapping of the S set onto the R set”. For example, in a task where the subject is required to perform a left- or a right-hand key press according to the lateral location of a visual stimulus, the ipsilateral S-R mapping instruction leads to shorter reaction times (RT) than the contralateral aer3. The difference in performance in favor of the fa-mer mapping is considered to be a typical instance of S-R compatibility. A number of RT studies have shown the additivity of the effect of S-R mapping with the respective effects of perceptus; and motor variables (e.g. ref. 6), which supports the idea that the effects of S-R compatibility can be assigned to the stage of S-R translation located between the perceptual and motor stages14.

Although, in the literature, emphasis has been placed on the influence of spatial S-R relationships, the effects of S-R compatibility are not restricted to the spatial correspondence between the respective elements of the S and R sets. indeed, what seems to be important for the emergence of an effect of S-R compatibility is the existence of a conceptual match between the S and R variables’. In recent formalizations, spatial S-R relationships are considered only as particular instances of the similarity between the subjective representations of the S and R variables’. In keeping with these formalizations, the present study was conducted in order to assess the compatibility relationship between the intensity of a sensory stimulus and the strength of a voluntary motor response. Several results suggest that both factors may affect not only the stage of S-R translation but also the subsequent motoric processes’*“. In these studies, overall RT was fractionated into two components with respect to the occurrence of the change in electromyographic

Correspondence: T. Hasbroucq, CNRS-LNC, 31 Chemin Joseph Aiguier, 13402 Marseille Cedex 09, France. Fax: (33) 91 22 08 75.

198 activity (EMG) of a muscle directly involved in the execution of the required response. The time interval between the onset of the stimulus snd the onset 0i EMG activity was termed ‘premotor time’, whilst the interval between the onset of EMG activity and the onset of the required response was termed ‘motor time’. The variati-ens of premotor time reflect the constraints of the central processes whereas variations of motor time reflect purely motoric processes. TO distinguish between the central locus of the effect of the manipulated variables and their potential motoric locus, the surface electromyographic activity of a prime mover of the thuntb, the adductor pollicis, has been recorded. Eight right-handed subjects (4 women and 4 men), aged 25-47 years (mean = 33). with normal or corrected-to-normal visual acuity volunteered for this experiment. The experiment took place in a darkened room. The subject sat at a table and gripped a manipulandum. The distal phalanx of the right thumb rested on a force sensor (24 mm in diameter) on which the subject had to produce a weak (from 2.45 to 15.44 Nl or a strong (fro;,1 15.46 to 63 Nl isometric press. The subject faced a white cardboard screen on which there was a vertical column of 4 light-emitting diodes (LED). The bottom most LED was positioned slightly below the line of gaze and about 50 cm away from the subject’s eyes. Switching this LED on served as the stimulus and its intensity, 0.216 mcd or 0.864 mcd, was varied by applying different currents. The upper series of three LEDs began 10 cm higher than the first one and they were each 9 mm apart. The pattern of illumination corresponded to three possible feedback messages: (I 1 when the force exerted was too low only the bottom LED was lit, (21 when the force exerted was in the correct range, the two inferior LEDs were lit, (3) when the force exerted was too strong, all three LEDs were lit. A trial began with the presentation of an auditory warning signal (2500 Hz, 75 dB, 50 ms in duration). After a delay of 1 s, one of the uvo stimuli (weak or strong) was presented and the response was to be executed during the following 2 s. After this interval, the stimulus was turned off and the feedback message was displayed for I s. The interval between the offset of the feedback message and the next warning signal was 1 s. Trials were presented in blocks of 32. Within a block, each possible stimulus (strong or weak) occurred I6 times in an unpredictable order. In each block, the first and second order sequential effects for trial-to-trial transitions were balanced12. There were two S-R mapping conditions: cornpa&

ble and incompatible. For the compatible mapping, the subjects were required to produce a strong press in response to the strong stimulus and a weak press in response to the weak stimulus. For the incompatible mapping, they had to produce a strong press in response to the weak stimulus and a weak press in response to the strong stimulus. There were three training sessions and one experimental session run on different days. The subjects were trained to respond correctly first in one mapping condition and then in the other during the first and second training sessions, respectively. The subjects were divided into two groups, A and B (2 men and 2 women in each group). Group A began training with the compatible condition and then had the incompatible condition in the following session. This order was reversed for group B. For each group, the criterion for the level of training in the first two sessions was attained when the subjects made no more than one error in each of two consecutive blocks. During the third training session, the mapping instruction was alternated every other block. In each group, one man and one woman started with the compatible mapping condition and the other subjects started with the incompatible mapping condition. All subjects succeeded in keeping their error rate below 5% during the experimental session. The experimental session during which the force and EMG signals were recorded was comprised of ten blocks and lasted about 1 h. The order of the conditions was the same as that adopted for the third training session. The thumb press was measured as a force signal using an Interlink Electronics FSR (Force Sensitive Resistor) and digitized on-line (A/D rate 1000 Hz). The EMG activity of the adductor pollicis muscle was recorded by means of paired surface Ag-AgCI electrodes, 12 mm in diameter and fixed about 1 cm apart. The EMG activity was amplified, filtered (high frequency cut-off 1 kHzl, full-wave rectified and digitized on-line (A/D rate 1000 Hz). For each trial, the force signal (solid line) and the EMG signal (dotted line) were displayed on the computer screen aligned to the onset of the imperative signal (Fig. 11. The onset of change in the force signal (force change, Fcl and the onset of change in the EMG signal (EMG change, EC) were determined after visual inspection and marked by placing a cursor with the computer mouse. A computer program determined the level and the time of peak force (respectively, force peak, Fp and force peak time, Fpt). Therefore, four chronometric indices cccld be estimated: the RT (measured from imperative signal to Fc), which was broken

199

down into a premotor time (from imperative signal to EC) and a motor time (from EC to Fc), and the contraction time was measured from Fc to Fpt. The angle (CU) between the segments Fc-Fp and Fc-Fpt was used as an index tto estimate the rate of change in force. The four chronometric indices (RT, premotor time, motor time and contraction time), the force peak and the cy angie were submitted to six separate repeatedmeasure canonical analyses of variance (ANOVA) with the mean square of the interaction between the effects of the subjects and the effect of the factor under analysis as the error term. The canonical ANOVAs involved two within-subject factors: signal intensity and response force. Error rates were compared across conditions using the non-parametric Wilcoxon Matched-Pairs SignedRanks T-test. The only statistically reliable difference in this dependent variable was that errors below the low criterion (below 2.45 N for the weak response or below 15.46 N for the strong one) were more numerous than errors above the high criterion (above 15.44 N for the weak response or above 63 N for the strong one); the error rates were 1.84% and 0.74%, respectively (T6 = 0, P s 0.05). RTs to the strong signal were shorter than RTs to the weak one (455 vs. 474 ms; F,,, = 8.74, P = 0.0212) and the strong response was produced faster than the weak one (451 vs. 478 ms; F,,., = 29, P = 0.001). Finally, the response to a strong signal was faster when it called for a strong response as opposed to a weak response: likewise, the response to a weak signal was faster when it called for a weak response as opposed to a strong

---m-f,_1 500

Time

from

,T__;” 1000

Imperative

-/ 1500

Signal

2000

(ms)

Fig. 1. The solid line represents the force signal and the dotted line the electromyographic (EMG) signal. The different indices recorded in this experiment were the reaction time CRT), the contraction time Cl?, the premotor time (PMTI. the motor time (MT), and a (see text for definitions). The other values marked on the figure are the onset of change in the EMG signal (EC), the onset of change in the force signal (Fc), the force peak (Fp), and the force peak time (Fpt).

-Weak

Imperative

strong Signal Intensity

Fig. 2. Premotor time (squares), motor time (triangles) and contraction time (circles) as a function of imperative signal intensity for the weak response (open symlzols) and for the strong response (fried symbols).

response !F,,, = 56.75, P = 0.0001). Therefore, the present results demonstrated a clear intensity to force compatibility effect on RT. The pattern of results for the premotor time closely mirrored that of RT (Fig. 2). Premotor time was shorter for the strong signal than for the weak one (383 vs. 400 ms; F,., = 7.91, P = 0.0261); it was also shorter for the strong response than for the weak one (451 vs. 478 ms; F,,, = 6.77, P = 0.0353). The Snteraction between these effects was highly significant (F,,, = 49.99, P = 0.0002). Again, this interaction indicated that premotor time was shorter when the imperative signal called for a compatible intensity to force mapping. In contrast, motor time was not affected by any of the factors manipulated. The effect of signal intensity, the effect of response force and the interaction between these effects were not statistically significant (respectively, F,,, = 1.46, P = 0.27; F,,, = 1.61, P = 0.25; F,,, = 2.42, P = 0.16). Thus, the effects of signal intensity and response force on overall RT were entirely due to premotoric processes. Contraction time (Fig. 2) was much longer for the strong response than for the weak response (295 ms vs. 189 ms; F,,, = 22, P = 0.0022). It was also longer for the weak imperative signal than for the strong one (246 vs. 238 ms; F,,, = 13.25, P = 0.0083). The interacticn between these effects, however, was non-significant (F,+, = 1.6, P = 0.32). The mean a-values were 58” for

and 72” for the strong response. regardless of the imperative signal intensity. This difference was statistically reliable (F,., = 29.27, P = 0.001). Neither the effect of the imperative signal intensity nor the interaction between this effect and that of the response force were significant (both Fs < 1). Therefore, the higher force required for the strong response, as compared to the weak one, was achieved through both a longer contraction time and a faster contraction. Force peak depended upon the required force (10.23 N vs. 46.12 N; F,.? = 630, p < O.ooOl) and upon stimulus intensity: it was higher for weak stimuli than for strong ones (28.78 vs. 27.57 N; F,,, = 7.72, P= 0.0274). The interaction between the effects of signal intensity and response force on force peak was non-significant (F < 1). Clearly, RT was shorter when the subjects were required to exert a strong press in response to the strong stimulus and a weak press in response to the weak stimulus than when they were required to perform the reverse S-R mapping. To the best of our knowledge, this is the first demonstration of an effect of intensity to force S-R compatibility. The analysis of the EMG signal revealed that this effect manifests itsell entirely before the generation of muscle action potentials in the prime mover studied (adductor pollicis). Thus, just as its more traditionally studied cognates, intensity to force compatibility is entirely attributable to premotoric processes. It is to be noted that the size of this new effect (69 ms) compares with that of the effect of spatial S-R mapping classically observed in two-choice RT experiments’“. In the literature, two types of S-R translation are generally distinguished, i.e. spatial transpositions and symbolic translations’h. Spatial transposition is when there is a physical spatial correspondence between the stimuli and the responses whereas symbolic translation is when the correspondence is not physical but learnt. The effects of spatial S-R compatibility have often been interpreted in terms of neuroanatomy. In many of the spatial tasks studied in the laboratory, the lateral, position of the stimuli and the responses in extra-corporal space correspond to the lateral organization of the neuraxis. When stimuli and responses are ipsilateral, they involve primary and motor areas of the same cerebral hemisphere but when stimuli are contralateral to the responses, then the primary sensory area of one hemisphere and the primary motor area of the other hemisphere are involved”. Several authors have attempted to explain the effects of spatial compatibility between stimulus and response in terms of interhemispheric transfer of information (cf. ref. 8) or in terms of asymmetric activation of the hemispheres (cf. ref. 5). the weak response

Although these hypotheses nowadays appear unjustified, in particular because they do not explain the effeds of spatial compatibility within the same hemispace (cf. ref. IS), a distinction is still made between spatial and symbolic compatibility in the literature (e.g. ref. 2). The effect of the intensity to force mapping demonstrated in the present study can be described perfectly in terms of the physical correspondence between stimulus and response. However, in contrast to the effects of spatial S-R compatibility, there is no neuroanatomical hypothesis that can explain the behavioral effect of this physical correspondence. A parsimonious afternative is that the effect is due to a symbolic transformation. Generalizing this alternative hypothesis to the effects of spatial S-R compatibility, we can consider that, iq agreement with Alluisi and Warm’, the accent placed on the role of the physical correspondence between stimulus and response is erroneous. We thank isahellr Mouret, Michel Bonnet, Jean Requin and JeanPierre Vedel for fruitful discussions. We also thank Raymond Fayolle and Guy Reynard for technical assistance. This research was supported by the Mini&e de la Recherche et de la Technologie (Grants 1991~co’),54 and 1991~CO923). the Centre National de la Recherche Scientifique (grant Cognisud)and the R&ion Provence Alpes C&e d’Azur (Grant 91/04324).

Alluisi, E.A. and Warm. J.S., Things that go together: a review of stimulus-response compatibility and related effects. In R.W. Proctor and T.G. Reeve (Eds.), Stimulus- Response Compatibiihy: An Inregrared fcrspecriw, North-Holland, Amsierdam. 1990, pp. 3-29. Bashore, T.R., Stimulus-response compatibility viewed from a cognitive psychophysiological perspective. In R.W. Proctor and T.G. Reeve (Eds. ). Stimuhrs - response Compatibility: AII Inregrarcd Perspcctil c, North-Holland, Amsterdam, 1990, pp. 183-223. Fitts. P.M. and Deininger. R.L.. S-R compatibility: correspondence among elements within stimulus and response codes. J. Erp, PsychoI., 4X ( l9S4) 483-492. Fit&, P.M. and Seeger. C.M., S-R compatibility: spatial characteristics of stimulus and respouse codes. J. Exp. Psychof., 46 (1953) 199-210. Guiard, Y.. Cerebral hemispheres Psycho/., 46 (1980) 41-61.

and selcl.; ~1‘ Lritention, ACM

6 Hasbroucq. T., Guiard, Y. and Kornblu..: :’ ‘P!.e additivity of stimulus-response compatibility with the c:! x:,s of sensory and motor factors in a tactile choice reaction time task, Acra Psycho/.. 72 ( 1989) 139- 144. 7 Hasbroucq, T., Guiard, Y. and Ottomani, L., Principles of response determination: the list-ru;e model of SR compatibility, &tll. Psychon. Sot., 28 f 199f I 327-330. 8 Hasbroucq, T.. Kornb!um. S. and Osman, A., A new look at reaction time estimates o: interhemispheric transmission time, Eur. Bull. Cogn. Psycho/.. 8 (1988) 207-221. 9 KawaL-, S.. Effects of force output and preparatory set on premotol ‘ime of simulta*leous bilateral responses, Percept. Mot Skills, 68 ( 1S,99) 619-6X. 10 Lute, R.D., Rcsponsc Times: Their Roles in Inferring Elementary MVUU~ Orgunizution, Oxford University Press, New York, 1986. 11 Poffenberger, A.T., Reaction time to retinal sdmulation with special reference to the time lost in conduction through nerve centers. Arch. Psycho/., 21 (1912) l-73.

201 12 Possama’i, C.-A. am! Reyaard. G.. SEREX 2: A program to generate exhaustive series, Behac: Res. Mcthds fns~r.. 6 (1974)

58. 13 Shaffer. L.H., Choice reaction with variable S-R mapping, 1. fip. psyclo~, 70 (1965) 244-288. 14 Stemberg, S.. The discovery of processing stages: Extensions of Donders’ method, Acru Psycl~ol.,30 (1969) 276-315.

15 Umiltk C. and Nkolrtri. R., Spatial stimulus-response compahbility, In R-W. Proctor and T.G. Reeve (Eds.), Srimtdus-mpmrp Cmpdbi/ir)‘: An htegrafed Prrspecfire, North-Holland. Amsterdam, 1990, pp. 89-116. I6 Wel ford, A.T., Ski&d Peflonnance: Perceptual and Motor SkiL!t, Scott Foreman, Glenview, 1976.