- Email: [email protected]
Decay of prism aftereffects under passive and active conditions
Cognitive Brain Research 20 (2004) 92 – 97 www.elsevier.com/locate/cogbrainres
Research report
Decay of prism aftereffects under passive and active conditions Juan Ferna´ndez-Ruiz a,b,*, Rosalinda Dı´az a, Carlos Aguilar a, Cynthia Hall-Haro a a
Departamento de Fisiologı´a, Facultad de Medicina, Universidad Nacional Auto´noma de Me´xico, Ciudad Universitaria CP 04510, Mexico City, D.F., Apartado Postal 70-250, Mexico b Escuela de Psicologı´a, Universidad Ana´huac, Mexico Accepted 22 January 2004 Available online 5 March 2004
Abstract In prism adaptation, subjects adapt to new visuospatial coordinates imposed by wedge prisms that laterally displace the visual field. During this process, subjects develop and store new visuomotor coordinates in order to compensate for the displacement of visual stimuli. After the prisms are removed, subjects show an aftereffect in the opposite direction of the original perturbation. The aftereffect is a manifestation of the recently stored information. In the present article, we were interested in studying the properties of the aftereffect. Specifically, we investigated the fate of the aftereffect under active conditions with motor reafferences but without visual input, and during passive conditions without visual or motor reafferences. The results in the motor active condition show that motor reafference (proprioceptive or corollary discharge information) led to a faster, but incomplete, aftereffect decay. The results in the passive condition show a bimodal aftereffect behavior, with a fast decay within the initial minutes, followed by a sustained aftereffect up to 20 min later. These data suggests that two different memory processes may contribute to the aftereffect, one showing a fast decay mainly within 1 min, and another that shows a stable endurance for more than 20 min. D 2004 Elsevier B.V. All rights reserved. Theme: Neural basis of behavior Topic: Learning and memory Keywords: Prism adaptation; Memory consolidation; Memory endurance; Memory decay; Visuomotor learning
1. Introduction When wedge prisms are donned, the induced visual perturbation produces an alteration of the normal visuomotor relationship. Prism adaptation refers to the modification of that relationship in order to acquaint for the perturbation induced by the prisms. The withdrawal of the prism during that process produces an aftereffect in the opposite direction of the original perturbation, proportional to the adaptation magnitude acquired before removing the prisms [4]. The whole prism adaptation process has been proposed to be a form of procedural learning (4), since it conforms nicely to Tulving’s definition of a procedural memory system as an action system whose operations are expressed * Corresponding author. Tel.: +52-55-56232123; fax: +52-5556363695. E-mail address: [email protected] (J. Ferna´ndez-Ruiz). 0926-6410/$ - see front matter D 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.cogbrainres.2004.01.007
in the form of skilled behavioral and cognitive procedures independent of any cognition [18]. It also falls within the boundaries of Squire’s nondeclarative learning definition that states that this kind of learning takes place when experience accumulates in behavioral change without affording conscious access to any memory content [16]. Viewed from this perspective, prism adaptation offers an important advantage for studying visuomotor learning, a specific kind of procedural learning, because it can dissociate performance from learning and memory. In a typical experiment, the behavioral modification elicited by the use of the prisms provides a learning estimate. Upon removal of the prisms, the aftereffect provides a measure of the persistence of the memory. Although there have been large efforts in order to understand prism adaptation [8,14,20], little is known about the fate of the new visuomotor map formed during this process. An important question that remains unre-
J. Ferna´ndez-Ruiz et al. / Cognitive Brain Research 20 (2004) 92–97
solved is if the new visuomotor mapping formed during prism adaptation decays spontaneously in passive conditions or remains unchanged ready to be used in the next visuomotor interaction. Previous reports addressing this issue have tested the same subjects at different time intervals after doffing the prisms [2,3,7,11 –13,17]. However, those studies did not tested the aftereffect decay under passive conditions because subjects still received proprioceptive or motor feedback (corollary discharge) information derived from the motor activity during each testing [1]. This pose two problems: first, after the initial aftereffect trial, the subsequent trials can be contaminated by cognitive factors because the subjects realise that they err even without prisms [19]; and second, and most important, since the adaptation – readaptation process is dependent on the number of visuomotor interactions [4], the observed decay at different intervals could have been the result of the accumulated interactions used to test previous delays (i.e. the test at 15 min was preceded by the tests at 5 or 10 min). It has been demonstrated in prism adaptation experiments that muscular proprioceptive information plays a role in the acquisition and retrieval of motor memory. For example faster aftereffect extinctions are found if muscular load or velocity are matched between the adaptation and aftereffect phases [5,6]. However, it is not currently known if motor reafferences, without visual reafferences, are capable of affecting the extinction of the new visuomotor mapping [1]. For this reason in the present study, we tested the influence of motor activity, similar to the one used during prism adaptation, on the decay function of the aftereffect. The results obtained suggested that even without visual feedback, the more motor activity, the faster the aftereffect decay. These results granted testing the passive decay function at different time intervals but using independent groups. The results obtained from the passive decay, without any kind of reafference, suggested the possibility to divide the aftereffect decay in two phases: one, lasting only a couple of minutes that shows a fast 40% decay, and a second phase, showing no decay at all, lasting at least 20 min.
2. Experiment 1 Since it is possible that motor reafferent stimulation has a role in the aftereffect decay, the first experiment was designed to test if execution of movements similar to those used during the acquisition, but without visual feedback, had some effect on the aftereffect extinction. 2.1. Materials and methods 2.1.1. Subjects Forty right-handed healthy subjects between the ages of 18 and 24 participated as volunteers in this experiment. Half of the subjects were female and the other half was male. The
93
subjects were naive to the purpose of the experiment and gave informed consent to participate prior to the experiments in accordance with the Declaration of Helsinki. 2.1.2. Basic prism adaptation paradigm We followed the prism adaptation throwing technique previously described [4]. Subjects threw clay balls (weight: 10 g) to a 12 12 cm cross drawn on a large sheet of parcel paper centered at shoulder level and placed 2 m away from them. The subjects were instructed to make each toss overhand during the whole experiment, and were asked to throw the balls to the location where they saw the target. The subjects performed the task from a standing position and had an unobstructed view of the target during the entire session. The head was unrestrained, and no directions were given about trunk, shoulder, or head/neck posture. However, they were not allowed to look down at their hand as they collected the next ball from a tray located right next to their bodies. A baseline throwing motor performance was obtained by having the subjects throw 26 balls at the target previous to the donning of the prisms (condition PRE). The position at which the balls made an impact on or around the target was marked immediately after each throw. After donning 30-diopter prisms, the subjects were instructed to throw 26 more balls in the same way (condition PRI; adaptation phase). 2.1.3. POS manipulation Before beginning the experiment, subjects were divided into two groups of 20 subjects each. Both groups followed the same PRE and PRI conditions previously described. After finishing their last PRI throw, they were asked to close their eyes before the post-prism testing (POS condition) began. During the POS conditions, after doffing the prisms, the first group was asked to remain with their eyes closed while making 13 throws to where they had seen the target previously, before opening their eyes to continue making 13 more throws to the, now visible, target. The second group followed the same instructions but instead of doing 13 throws, they made 26 throws with their eyes closed before they continue making 13 more throws to the visible target. During the throws made with the closed eyes, the subjects were asked to remain stood without making major body movements, just as they were while making the PRI throws that preceded the POS phase. Three measures were calculated from the collected data. First, an adaptation measure was obtained by subtracting the distance to the center of the ball’s impact on the final throw while wearing the prisms, from that on the initial throw while wearing them. Second, a POS with eyes closed condition was the measurement of the ball’s impact to the target on the first throw with the eyes closed after removing the prisms. And the third measurement was the aftereffect that was defined as the ball’s impact horizontal distance to the target on the first throw with the eyes open after testing with the eyes closed.
94
J. Ferna´ndez-Ruiz et al. / Cognitive Brain Research 20 (2004) 92–97
2.1.4. Statistical analysis Student’s t-tests were used to compare independent groups. To compare data within groups, paired t-tests were used. 2.2. Results and discussion The raw data obtained in this experiment is shown in Fig. 1A and B. The horizontal mean distance from the target in the POS condition with the eyes closed suggest that subjects in both groups were capable of centering their throws around the target, even though they were not looking at it (x = 0. 68 + 0.73 and 0.15 + 0.53). A Student’s t-test showed that there were no differences between both groups (t = 0.576, DF = 37; p = 0.56 ns).
However, the most relevant results are those concerning the aftereffect, once the subjects were instructed to open their eyes again. A paired t-test comparison of the last PRE baseline throw vs. the first POS throw with eyes open showed significant differences in both groups, suggesting a preservation of the acquired visuomotor map (t = 6.39, DF = 19; p < 0.001 for the 13 throws group; t = 4.86, DF = 19; p < 0.01 for the 26 throws group). Fig. 1C shows the first throw with the eyes closed and the first throw with the eyes open during POS for both groups. A comparison of the first throw with eyes open in the POS condition between both groups revealed a significant decay of the 26 throws vs. the 13 throws group (t = 2.22, DF = 38; p = 0.03). These results suggest that even though the motor reafference by itself does not completely abolish the aftereffect it does produce a faster decay.
3. Experiment 2 The first experiment suggests that motor activity, similar to the one used when adapting, lead to faster aftereffect extinction. Since previous studies of the aftereffect decay function tested the same subjects at different intervals, it could be possible that their results were contaminated by the subjects’ activity during testing. In this experiment, we studied the aftereffect decay function in independent groups under passive conditions without visual or motor reafferences. 3.1. Materials and methods 3.1.1. Subjects One hundred right-handed healthy subjects between the ages of 18 and 24 participated as volunteers in this study. Half of the subjects were female and the other half was male. The subjects were naive to the purpose of the experiment and gave informed consent to participate prior to the experiments in accordance with the Declaration of Helsinki.
Fig. 1. Raw data obtained in conditions PRE, PRI and POS for the 13 throws group (A) and the 26 throws group (B). Circles depict throws made with the eyes open, and rhombus depict throws made with the eyes closed; arrows indicate the first POS throw with the eyes open; (C) shows the POS (aftereffect) data in absolute values for the first throw with the eyes closed (black bars) and the first throw with the eyes open (gray bars) for both groups. *Denotes significant difference at p < 0.05. Error bars = S.E.M.
3.1.2. Basic prism adaptation paradigm We followed the same basic paradigm, except that after removing the prisms used in the PRI condition, the subjects threw 26 more balls to the target with the eyes open (condition POS). Three measures were calculated from the collected data. First, an adaptation measure was obtained by subtracting the distance to the center of the ball’s impact on the final throw while wearing the prisms, from that on the initial throw while wearing them. Second, an aftereffect measure was defined as the ball’s impact horizontal distance to the target on the first throw after removing the prisms. Third, an aftereffect as a percentage of the previous adaptation was obtained. This last measurement is necessary because each group can have different adaptation magnitudes, and since
J. Ferna´ndez-Ruiz et al. / Cognitive Brain Research 20 (2004) 92–97
the aftereffect is directly correlated with the adaptation, then a transformation of the raw aftereffect data has to be made into a percentage of the adaptation. For example, it could happen that two groups have the same aftereffect magnitude, i.e. of 10 cm. But one had an adaptation magnitude of 20 cm, while the other had an adaptation magnitude of 40 cm. If we did not apply the transformation, an aftereffect comparison would show no differences, while a comparison of the aftereffect as a percentage of the adaptation would show the real group differences of 25% vs. 50% of the adaptation. 3.1.3. Delay Subjects were divided into five groups of 20 subjects each. Each group was assigned randomly to a single delay that could be of 0.1, 1, 5, 10 or 20 min. The delay period was introduced between the last PRI throw, and the first POS throw. During delay, the room was completely dark, and the subjects were asked to stay relaxed with their eyes closed. 3.1.4. Statistical analysis In order to compare the different delay groups, a Kruskal – Wallis One Way Analysis of Variance on Ranks was made. This method was chosen because the analysis was made on transformed data (see above), and not on the raw data. An all Pairwise Multiple Comparison Procedures (Student– Newman – Keuls Method) were made to know the specific differences among groups. The extinction rate was analyzed by finding the best fit to the averaged data, and then obtaining the derivatives at different time intervals. 3.2. Results and discussion Fig. 2 shows the aftereffect at different delays as a percentage of the aftereffect of the 0.1 delay group. Having a delay between the last throw while wearing the prism and
95
the first throw after doffing them had a significant impact on the aftereffect magnitude. In fact a Kruskal – Wallis one way ANOVA on Ranks showed a statistical difference among groups (H = 9.95, DF = 4; p < 0.05). A subsequent Student – Newman –Keuls all pairwise multiple comparison showed that the 0.1 delay group was different from all the other groups. No other significant differences were found in the analysis. In order to know if delay reduced the aftereffect to levels similar to the previous baseline magnitude, a paired t-test comparison between each group’s aftereffect and their previous baseline was done. The analysis demonstrated that in all the groups, the aftereffect magnitude was significantly different from their previous baseline (all p’s < 0.01). A one way ANOVA on the baseline (PRE) last throw demonstrated that there were no baseline differences among groups ( F(4,95) = 7; p = 0.59). To analyze the extinction, the data was fitted to a curve using specialised software (CurveExpert 1.3, by Daniel Hyams, Starkville, MS 39759 USA). A Power Fit ( y = axb) was found to had the best fit (r = 0.95) on the decay data (Fig. 2). The analysis of the regression showed that the initial decay from 0.1 to 1 min had the major impact on the aftereffect extinction since the derivatives went from 13.5 to 8.3. The subsequent rate of change was much smaller, showing a derivative of 0.5, 0.5 and 0.04 for the measurements made at 5, 10 and 20 min after doffing the prisms. After the initial minute, the aftereffect decay was 25%, and by 10 min 40% decay had accumulated. In contrast, the second 10 min (from min. 10 to min. 20) were characterized by a steady state with almost no extinction rate. The first throw with the eyes open for both groups of the first experiment were compared to the aftereffect of the different groups of the second experiment. The results show that the 13 throws group is similar to the 1 min delay group of the first experiment, while the 26 throws group is similar to the 20 min delay group (t = 1.67, DF = 38; p = 0.1 ns); since making 13 throws takes less than a minute, the difference can only be explained by a faster decay in the 26 throws group compared to the 13 throws group.
4. General discussion
Fig. 2. Aftereffect at different delays as a percentage of the aftereffect of the 0.1 delay group. The line represents the best fit using a power function. The only significant difference found was between delay 0.1 and all other groups ( p < 0.05).
The present study addresses two important questions pertaining procedural memory in general, and prism adaptation in particular. First, in the absence of visual reafference, does motor reafference information from motor commands or proprioceptive information affects the spontaneous decay observed in passive conditions? And second, does newly form visuomotor mappings decay spontaneously in the complete absence of reafference stimulation? The results obtained in the two experiments suggest that active motor reafference, in the absence of visual feedback, produces faster aftereffect decay. The results also suggest that there is a fast aftereffect extinction component which occurs
96
J. Ferna´ndez-Ruiz et al. / Cognitive Brain Research 20 (2004) 92–97
under passive conditions, followed by a more enduring component that last more than 20 min. Following is a detailed discussion of these two aspects. The effect of active motor reafference in the absence of visual feedback is an important question, since previous reports tested the same subjects at different time intervals. Although the subjects were deprived of visual feedback (by turning off lights, or by occluding vision) the fact that they were making the same movements was not acquainted as a possible variable [2,3,7,17]. Here we have shown that motor activity does produce a faster, but incomplete, aftereffect decay. It has been proposed that the aftereffect is the result of the contribution of different components [2,5]. If, for example, proprioception contributes to some extent to the total aftereffect magnitude, then the active movements could specifically contribute to a faster extinction of that component, resulting in a faster but partial, decay of the whole aftereffect measurement [2]. A modular decay has been proven in prism adaptation using this and other models [5,6]. However, at this point we have not investigated the possible contribution of different motor information sources like corollary discharge activity, or proprioceptive information. The aftereffect decay in the prism adaptation paradigm has been previously investigated. In an early attempt to resolve this question, Hamilton and Bossom [22] found a 50% aftereffect reduction in subjects that sat in the dark for 15 min. They tested if the same conditions to establish prism altered visuomotor mappings were required for returning to the original mapping, after prism adaptation had taken place. They specifically tested if visual reafferent stimulation was necessary for the return of the normal coordination. After their subjects sat in a dark room for 15 min following adaptation, they were tested again without prisms. The authors found a significant reduction of the aftereffect, suggesting to them that the normal visuomotor mapping is not lost, but is retained and reinstated following prism removal, even in the absence of any visual feedback. In the present article, the results from the second experiment suggest that although there is a passive extinction, a large component of the aftereffect (60%) persists over 20 min after training. In another study, Taub and Goldberg [17] show decay functions for aftereffects during 60 min after prism removal. The authors report that after 15 min subjects still had around 40% aftereffects in average, but after 1 h of testing almost every 5 min, only one of their groups showed around one fourth of the aftereffect. These results contrast with those by Dewar [3], who report a much larger aftereffect persistence, around 75% after 15 min. Choe and Welch [2] also found a 50 –60% decay of the aftereffect after 15 min, with continuous exposure groups decaying faster than terminal exposure groups. However, the studies that obtained decay functions of the aftereffect tested the same subjects at different time intervals. The present results provide the first function of the aftereffect extinction rate not affected by visual or motor reafferences.
Our results suggest that the aftereffect has two components: one that shows a fast extinction in the absence of further visuomotor interactions, and another that shows a longer endurance. The idea of having two components with different time ranges has been previously advanced for explaining changes in the timing of eye – hand coordination during prism adaptation [15]. It has been argued that the former component can be strategically used to quickly correct misalignment, while the later component would be responsible for slower long-term alignments [15]. The data showed in the present article suggest that two different memory processes may contribute to the aftereffect magnitude, one showing a fast decay mainly within 1 min, and another that shows a stable endurance for more than 20 min. The fast decaying component could be related to motor working memory that depends on factors like attention, while the second, more stable component could hold information acquired through procedural learning mechanisms. It is interesting to note that although in a typical prism adaptation/readaptation experiment the adaptation disappears rapidly after removing the prisms, the continuous training with prisms across several days or weeks result in the acquisition of a second or even more visuomotor correlations that can be accessed in the long term [9,10,21]. The consolidation of a second long-term visuomotor correlation suggests that after each complete adaptation/readaptation process some information remains, even if the subject had already went back to the original baseline visuomotor performance. We propose that the long-term component reported in the present experiment is different from the long-term formation of a second visuomotor calibration. One important difference would be that the long-term component we are describing tend to disappear as soon as the subject starts a visuomotor interaction without the prisms. In contrast, subjects that have acquired a longterm second visuomotor calibration can quickly switch between the old and the new calibrations depending on different factors like the presence or absence of the prisms frames [9]. However, it could be possible that some information stored during the second component described in this article could contribute to the acquisition of a long-term visuomotor correlation of the type obtained through extensive training across many sessions. It has been previously shown that under normal visuomotor reafference conditions complete aftereffect decay can be observed in less than 1 min of continuous interactions. Here we have shown that increasing the number of motor reafferences of the same nature as those used during training can accelerate the aftereffect decay in the absence of visual feedback. We also showed that under passive conditions it is possible to observe two different phases within the aftereffect. An initial period is characterized by a fast decay of about 40% of the total magnitude followed by a second period in which there is almost no extinction of the aftereffect.
J. Ferna´ndez-Ruiz et al. / Cognitive Brain Research 20 (2004) 92–97
Acknowledgements We thank Silvia Revuelta and Rafael Ojeda for their help in testing the subjects. This work was supported by CONACyT 34817-M, 30970-M and DGAPA IN210300.
[10] [11] [12]
References [1] J.S. Baizer, I. Kralj-Hans, M. Glickstein, Cerebellar lesions and prism adaptation in macaque monkeys, J. Neurophysiol. 81 (1999) 1960 – 1965. [2] C.S. Choe, R.B. Welch, Variables affecting the intermanual transfer and decay of prism adaptation, J. Exp. Psychol. 102 (1974) 1076 – 1084. [3] R. Dewar, Adaptation to displaced vision—the influence of distribution of practice on retention, Percept. Psychophys. 8 (1970) 33 – 34. [4] J. Fernandez-Ruiz, R. Diaz, Prism adaptation and aftereffect: specifying the properties of a procedural memory system, Learn. Mem. 6 (1999) 47 – 53. [5] J. Fernandez-Ruiz, C. Hall-Haro, R. Diaz, J. Mischner, P. Vergara, J.C. Lopez-Garcia, Learning motor synergies makes use of information on muscular load, Learn. Mem. 7 (2000) 193 – 198. [6] S. Kitazawa, T. Kimura, T. Uka, Prism adaptation of reaching movements: specificity for the velocity of reaching, J. Neurosci. 17 (1997) 1481 – 1492. [7] S.T. Klapp, S.A. Nordell, K.C. Hoekenga, C.B. Patton, Long-lasting aftereffect of brief prism exposure, Percept. Psychophys. 15 (1974) 399 – 400. [8] A.S. Kornheiser, Adaptation to laterally displaced vision: a review, Psychol. Bull. 83 (1976) 783 – 816. [9] T.A.Martin,J.G.Keating,H.P.Goodkin,A.J.Bastian,W.T.Thach,Thach, Thach, Throwing while looking through prisms: II. Specificity and
[13]
[14] [15]
[16]
[17] [18]
[19]
[20] [21]
[22]
97
storage of multiple gaze-throw calibrations, Brain 119 (Pt 4) (1996) 1199 – 1211. B.O. McGonigle, J. Flook, Long-term retention of single and multistate prismatic adaptation by humans, Nature 272 (1978) 364 – 366. G.M. Redding, Visual adaptation to tilt and displacement: same or different processes? Percept. Psychophys. 14 (1973) 193 – 200. G.M. Redding, Decay of visual adaptation to tilt and displacement, Percept. Psychophys. 17 (1975) 203 – 208. G.M. Redding, B. Wallace, Components of displacement adaptation in acquisition and decay as a function of hand and hall exposure, Percept. Psychophys. 20 (1976) 453 – 459. G.M. Redding, B. Wallace, Adaptive Spatial Alignment, vol. x, Lawrence Erlbaum Associates, Mahwah, NJ, 1997 194 pp. Y.K.K. Rossetti, T. Mano, Prismatic displacement of vision induces transient changes in the timing of eye – hand coordination, Percept. Psychophys. 54 (1993) 355 – 364. L. Squire, Declarative and nondeclarative memory: multiple brain systems supporting learning and memory, in: P. Andersen (Ed.), Memory Concepts. Basic and Clinical Aspects, Excerpta Medica, New York, 1993, pp. 3 – 25. E. Taub, L.A. Goldberg, Prism adaptation: control of intermanual transfer by distribution of practice, Science 180 (1973) 755 – 757. E. Tulving, Human memory, in: P. Andersen (Ed.), Memory Concepts. Basic and Clinical Aspects, Excerpta Medica, New York, 1993, pp. 27 – 45. M.J. Weiner, M. Hallett, H.H. Funkenstein, Adaptation to lateral displacement of vision in patients with lesions of the central nervous system, Neurology 33 (1983) 766 – 772. R.B. Welch, Perceptual Modification: Adapting to Altered Sensory Environments, vol. xv, Academic Press, New York, 1978 346 pp. R.B. Welch, B. Bridgeman, S. Anand, K.E. Browman, Alternating prism exposure causes dual adaptation and generalization to a novel displacement, Percept. Psychophys. 54 (1993) 195 – 204. C.R. Hamilton, J. Bossom, Decay of prism aftereffects, J. Exp. Psychol. 67 (1964) 148 – 150.
Research report
Decay of prism aftereffects under passive and active conditions Juan Ferna´ndez-Ruiz a,b,*, Rosalinda Dı´az a, Carlos Aguilar a, Cynthia Hall-Haro a a
Departamento de Fisiologı´a, Facultad de Medicina, Universidad Nacional Auto´noma de Me´xico, Ciudad Universitaria CP 04510, Mexico City, D.F., Apartado Postal 70-250, Mexico b Escuela de Psicologı´a, Universidad Ana´huac, Mexico Accepted 22 January 2004 Available online 5 March 2004
Abstract In prism adaptation, subjects adapt to new visuospatial coordinates imposed by wedge prisms that laterally displace the visual field. During this process, subjects develop and store new visuomotor coordinates in order to compensate for the displacement of visual stimuli. After the prisms are removed, subjects show an aftereffect in the opposite direction of the original perturbation. The aftereffect is a manifestation of the recently stored information. In the present article, we were interested in studying the properties of the aftereffect. Specifically, we investigated the fate of the aftereffect under active conditions with motor reafferences but without visual input, and during passive conditions without visual or motor reafferences. The results in the motor active condition show that motor reafference (proprioceptive or corollary discharge information) led to a faster, but incomplete, aftereffect decay. The results in the passive condition show a bimodal aftereffect behavior, with a fast decay within the initial minutes, followed by a sustained aftereffect up to 20 min later. These data suggests that two different memory processes may contribute to the aftereffect, one showing a fast decay mainly within 1 min, and another that shows a stable endurance for more than 20 min. D 2004 Elsevier B.V. All rights reserved. Theme: Neural basis of behavior Topic: Learning and memory Keywords: Prism adaptation; Memory consolidation; Memory endurance; Memory decay; Visuomotor learning
1. Introduction When wedge prisms are donned, the induced visual perturbation produces an alteration of the normal visuomotor relationship. Prism adaptation refers to the modification of that relationship in order to acquaint for the perturbation induced by the prisms. The withdrawal of the prism during that process produces an aftereffect in the opposite direction of the original perturbation, proportional to the adaptation magnitude acquired before removing the prisms [4]. The whole prism adaptation process has been proposed to be a form of procedural learning (4), since it conforms nicely to Tulving’s definition of a procedural memory system as an action system whose operations are expressed * Corresponding author. Tel.: +52-55-56232123; fax: +52-5556363695. E-mail address: [email protected] (J. Ferna´ndez-Ruiz). 0926-6410/$ - see front matter D 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.cogbrainres.2004.01.007
in the form of skilled behavioral and cognitive procedures independent of any cognition [18]. It also falls within the boundaries of Squire’s nondeclarative learning definition that states that this kind of learning takes place when experience accumulates in behavioral change without affording conscious access to any memory content [16]. Viewed from this perspective, prism adaptation offers an important advantage for studying visuomotor learning, a specific kind of procedural learning, because it can dissociate performance from learning and memory. In a typical experiment, the behavioral modification elicited by the use of the prisms provides a learning estimate. Upon removal of the prisms, the aftereffect provides a measure of the persistence of the memory. Although there have been large efforts in order to understand prism adaptation [8,14,20], little is known about the fate of the new visuomotor map formed during this process. An important question that remains unre-
J. Ferna´ndez-Ruiz et al. / Cognitive Brain Research 20 (2004) 92–97
solved is if the new visuomotor mapping formed during prism adaptation decays spontaneously in passive conditions or remains unchanged ready to be used in the next visuomotor interaction. Previous reports addressing this issue have tested the same subjects at different time intervals after doffing the prisms [2,3,7,11 –13,17]. However, those studies did not tested the aftereffect decay under passive conditions because subjects still received proprioceptive or motor feedback (corollary discharge) information derived from the motor activity during each testing [1]. This pose two problems: first, after the initial aftereffect trial, the subsequent trials can be contaminated by cognitive factors because the subjects realise that they err even without prisms [19]; and second, and most important, since the adaptation – readaptation process is dependent on the number of visuomotor interactions [4], the observed decay at different intervals could have been the result of the accumulated interactions used to test previous delays (i.e. the test at 15 min was preceded by the tests at 5 or 10 min). It has been demonstrated in prism adaptation experiments that muscular proprioceptive information plays a role in the acquisition and retrieval of motor memory. For example faster aftereffect extinctions are found if muscular load or velocity are matched between the adaptation and aftereffect phases [5,6]. However, it is not currently known if motor reafferences, without visual reafferences, are capable of affecting the extinction of the new visuomotor mapping [1]. For this reason in the present study, we tested the influence of motor activity, similar to the one used during prism adaptation, on the decay function of the aftereffect. The results obtained suggested that even without visual feedback, the more motor activity, the faster the aftereffect decay. These results granted testing the passive decay function at different time intervals but using independent groups. The results obtained from the passive decay, without any kind of reafference, suggested the possibility to divide the aftereffect decay in two phases: one, lasting only a couple of minutes that shows a fast 40% decay, and a second phase, showing no decay at all, lasting at least 20 min.
2. Experiment 1 Since it is possible that motor reafferent stimulation has a role in the aftereffect decay, the first experiment was designed to test if execution of movements similar to those used during the acquisition, but without visual feedback, had some effect on the aftereffect extinction. 2.1. Materials and methods 2.1.1. Subjects Forty right-handed healthy subjects between the ages of 18 and 24 participated as volunteers in this experiment. Half of the subjects were female and the other half was male. The
93
subjects were naive to the purpose of the experiment and gave informed consent to participate prior to the experiments in accordance with the Declaration of Helsinki. 2.1.2. Basic prism adaptation paradigm We followed the prism adaptation throwing technique previously described [4]. Subjects threw clay balls (weight: 10 g) to a 12 12 cm cross drawn on a large sheet of parcel paper centered at shoulder level and placed 2 m away from them. The subjects were instructed to make each toss overhand during the whole experiment, and were asked to throw the balls to the location where they saw the target. The subjects performed the task from a standing position and had an unobstructed view of the target during the entire session. The head was unrestrained, and no directions were given about trunk, shoulder, or head/neck posture. However, they were not allowed to look down at their hand as they collected the next ball from a tray located right next to their bodies. A baseline throwing motor performance was obtained by having the subjects throw 26 balls at the target previous to the donning of the prisms (condition PRE). The position at which the balls made an impact on or around the target was marked immediately after each throw. After donning 30-diopter prisms, the subjects were instructed to throw 26 more balls in the same way (condition PRI; adaptation phase). 2.1.3. POS manipulation Before beginning the experiment, subjects were divided into two groups of 20 subjects each. Both groups followed the same PRE and PRI conditions previously described. After finishing their last PRI throw, they were asked to close their eyes before the post-prism testing (POS condition) began. During the POS conditions, after doffing the prisms, the first group was asked to remain with their eyes closed while making 13 throws to where they had seen the target previously, before opening their eyes to continue making 13 more throws to the, now visible, target. The second group followed the same instructions but instead of doing 13 throws, they made 26 throws with their eyes closed before they continue making 13 more throws to the visible target. During the throws made with the closed eyes, the subjects were asked to remain stood without making major body movements, just as they were while making the PRI throws that preceded the POS phase. Three measures were calculated from the collected data. First, an adaptation measure was obtained by subtracting the distance to the center of the ball’s impact on the final throw while wearing the prisms, from that on the initial throw while wearing them. Second, a POS with eyes closed condition was the measurement of the ball’s impact to the target on the first throw with the eyes closed after removing the prisms. And the third measurement was the aftereffect that was defined as the ball’s impact horizontal distance to the target on the first throw with the eyes open after testing with the eyes closed.
94
J. Ferna´ndez-Ruiz et al. / Cognitive Brain Research 20 (2004) 92–97
2.1.4. Statistical analysis Student’s t-tests were used to compare independent groups. To compare data within groups, paired t-tests were used. 2.2. Results and discussion The raw data obtained in this experiment is shown in Fig. 1A and B. The horizontal mean distance from the target in the POS condition with the eyes closed suggest that subjects in both groups were capable of centering their throws around the target, even though they were not looking at it (x = 0. 68 + 0.73 and 0.15 + 0.53). A Student’s t-test showed that there were no differences between both groups (t = 0.576, DF = 37; p = 0.56 ns).
However, the most relevant results are those concerning the aftereffect, once the subjects were instructed to open their eyes again. A paired t-test comparison of the last PRE baseline throw vs. the first POS throw with eyes open showed significant differences in both groups, suggesting a preservation of the acquired visuomotor map (t = 6.39, DF = 19; p < 0.001 for the 13 throws group; t = 4.86, DF = 19; p < 0.01 for the 26 throws group). Fig. 1C shows the first throw with the eyes closed and the first throw with the eyes open during POS for both groups. A comparison of the first throw with eyes open in the POS condition between both groups revealed a significant decay of the 26 throws vs. the 13 throws group (t = 2.22, DF = 38; p = 0.03). These results suggest that even though the motor reafference by itself does not completely abolish the aftereffect it does produce a faster decay.
3. Experiment 2 The first experiment suggests that motor activity, similar to the one used when adapting, lead to faster aftereffect extinction. Since previous studies of the aftereffect decay function tested the same subjects at different intervals, it could be possible that their results were contaminated by the subjects’ activity during testing. In this experiment, we studied the aftereffect decay function in independent groups under passive conditions without visual or motor reafferences. 3.1. Materials and methods 3.1.1. Subjects One hundred right-handed healthy subjects between the ages of 18 and 24 participated as volunteers in this study. Half of the subjects were female and the other half was male. The subjects were naive to the purpose of the experiment and gave informed consent to participate prior to the experiments in accordance with the Declaration of Helsinki.
Fig. 1. Raw data obtained in conditions PRE, PRI and POS for the 13 throws group (A) and the 26 throws group (B). Circles depict throws made with the eyes open, and rhombus depict throws made with the eyes closed; arrows indicate the first POS throw with the eyes open; (C) shows the POS (aftereffect) data in absolute values for the first throw with the eyes closed (black bars) and the first throw with the eyes open (gray bars) for both groups. *Denotes significant difference at p < 0.05. Error bars = S.E.M.
3.1.2. Basic prism adaptation paradigm We followed the same basic paradigm, except that after removing the prisms used in the PRI condition, the subjects threw 26 more balls to the target with the eyes open (condition POS). Three measures were calculated from the collected data. First, an adaptation measure was obtained by subtracting the distance to the center of the ball’s impact on the final throw while wearing the prisms, from that on the initial throw while wearing them. Second, an aftereffect measure was defined as the ball’s impact horizontal distance to the target on the first throw after removing the prisms. Third, an aftereffect as a percentage of the previous adaptation was obtained. This last measurement is necessary because each group can have different adaptation magnitudes, and since
J. Ferna´ndez-Ruiz et al. / Cognitive Brain Research 20 (2004) 92–97
the aftereffect is directly correlated with the adaptation, then a transformation of the raw aftereffect data has to be made into a percentage of the adaptation. For example, it could happen that two groups have the same aftereffect magnitude, i.e. of 10 cm. But one had an adaptation magnitude of 20 cm, while the other had an adaptation magnitude of 40 cm. If we did not apply the transformation, an aftereffect comparison would show no differences, while a comparison of the aftereffect as a percentage of the adaptation would show the real group differences of 25% vs. 50% of the adaptation. 3.1.3. Delay Subjects were divided into five groups of 20 subjects each. Each group was assigned randomly to a single delay that could be of 0.1, 1, 5, 10 or 20 min. The delay period was introduced between the last PRI throw, and the first POS throw. During delay, the room was completely dark, and the subjects were asked to stay relaxed with their eyes closed. 3.1.4. Statistical analysis In order to compare the different delay groups, a Kruskal – Wallis One Way Analysis of Variance on Ranks was made. This method was chosen because the analysis was made on transformed data (see above), and not on the raw data. An all Pairwise Multiple Comparison Procedures (Student– Newman – Keuls Method) were made to know the specific differences among groups. The extinction rate was analyzed by finding the best fit to the averaged data, and then obtaining the derivatives at different time intervals. 3.2. Results and discussion Fig. 2 shows the aftereffect at different delays as a percentage of the aftereffect of the 0.1 delay group. Having a delay between the last throw while wearing the prism and
95
the first throw after doffing them had a significant impact on the aftereffect magnitude. In fact a Kruskal – Wallis one way ANOVA on Ranks showed a statistical difference among groups (H = 9.95, DF = 4; p < 0.05). A subsequent Student – Newman –Keuls all pairwise multiple comparison showed that the 0.1 delay group was different from all the other groups. No other significant differences were found in the analysis. In order to know if delay reduced the aftereffect to levels similar to the previous baseline magnitude, a paired t-test comparison between each group’s aftereffect and their previous baseline was done. The analysis demonstrated that in all the groups, the aftereffect magnitude was significantly different from their previous baseline (all p’s < 0.01). A one way ANOVA on the baseline (PRE) last throw demonstrated that there were no baseline differences among groups ( F(4,95) = 7; p = 0.59). To analyze the extinction, the data was fitted to a curve using specialised software (CurveExpert 1.3, by Daniel Hyams, Starkville, MS 39759 USA). A Power Fit ( y = axb) was found to had the best fit (r = 0.95) on the decay data (Fig. 2). The analysis of the regression showed that the initial decay from 0.1 to 1 min had the major impact on the aftereffect extinction since the derivatives went from 13.5 to 8.3. The subsequent rate of change was much smaller, showing a derivative of 0.5, 0.5 and 0.04 for the measurements made at 5, 10 and 20 min after doffing the prisms. After the initial minute, the aftereffect decay was 25%, and by 10 min 40% decay had accumulated. In contrast, the second 10 min (from min. 10 to min. 20) were characterized by a steady state with almost no extinction rate. The first throw with the eyes open for both groups of the first experiment were compared to the aftereffect of the different groups of the second experiment. The results show that the 13 throws group is similar to the 1 min delay group of the first experiment, while the 26 throws group is similar to the 20 min delay group (t = 1.67, DF = 38; p = 0.1 ns); since making 13 throws takes less than a minute, the difference can only be explained by a faster decay in the 26 throws group compared to the 13 throws group.
4. General discussion
Fig. 2. Aftereffect at different delays as a percentage of the aftereffect of the 0.1 delay group. The line represents the best fit using a power function. The only significant difference found was between delay 0.1 and all other groups ( p < 0.05).
The present study addresses two important questions pertaining procedural memory in general, and prism adaptation in particular. First, in the absence of visual reafference, does motor reafference information from motor commands or proprioceptive information affects the spontaneous decay observed in passive conditions? And second, does newly form visuomotor mappings decay spontaneously in the complete absence of reafference stimulation? The results obtained in the two experiments suggest that active motor reafference, in the absence of visual feedback, produces faster aftereffect decay. The results also suggest that there is a fast aftereffect extinction component which occurs
96
J. Ferna´ndez-Ruiz et al. / Cognitive Brain Research 20 (2004) 92–97
under passive conditions, followed by a more enduring component that last more than 20 min. Following is a detailed discussion of these two aspects. The effect of active motor reafference in the absence of visual feedback is an important question, since previous reports tested the same subjects at different time intervals. Although the subjects were deprived of visual feedback (by turning off lights, or by occluding vision) the fact that they were making the same movements was not acquainted as a possible variable [2,3,7,17]. Here we have shown that motor activity does produce a faster, but incomplete, aftereffect decay. It has been proposed that the aftereffect is the result of the contribution of different components [2,5]. If, for example, proprioception contributes to some extent to the total aftereffect magnitude, then the active movements could specifically contribute to a faster extinction of that component, resulting in a faster but partial, decay of the whole aftereffect measurement [2]. A modular decay has been proven in prism adaptation using this and other models [5,6]. However, at this point we have not investigated the possible contribution of different motor information sources like corollary discharge activity, or proprioceptive information. The aftereffect decay in the prism adaptation paradigm has been previously investigated. In an early attempt to resolve this question, Hamilton and Bossom [22] found a 50% aftereffect reduction in subjects that sat in the dark for 15 min. They tested if the same conditions to establish prism altered visuomotor mappings were required for returning to the original mapping, after prism adaptation had taken place. They specifically tested if visual reafferent stimulation was necessary for the return of the normal coordination. After their subjects sat in a dark room for 15 min following adaptation, they were tested again without prisms. The authors found a significant reduction of the aftereffect, suggesting to them that the normal visuomotor mapping is not lost, but is retained and reinstated following prism removal, even in the absence of any visual feedback. In the present article, the results from the second experiment suggest that although there is a passive extinction, a large component of the aftereffect (60%) persists over 20 min after training. In another study, Taub and Goldberg [17] show decay functions for aftereffects during 60 min after prism removal. The authors report that after 15 min subjects still had around 40% aftereffects in average, but after 1 h of testing almost every 5 min, only one of their groups showed around one fourth of the aftereffect. These results contrast with those by Dewar [3], who report a much larger aftereffect persistence, around 75% after 15 min. Choe and Welch [2] also found a 50 –60% decay of the aftereffect after 15 min, with continuous exposure groups decaying faster than terminal exposure groups. However, the studies that obtained decay functions of the aftereffect tested the same subjects at different time intervals. The present results provide the first function of the aftereffect extinction rate not affected by visual or motor reafferences.
Our results suggest that the aftereffect has two components: one that shows a fast extinction in the absence of further visuomotor interactions, and another that shows a longer endurance. The idea of having two components with different time ranges has been previously advanced for explaining changes in the timing of eye – hand coordination during prism adaptation [15]. It has been argued that the former component can be strategically used to quickly correct misalignment, while the later component would be responsible for slower long-term alignments [15]. The data showed in the present article suggest that two different memory processes may contribute to the aftereffect magnitude, one showing a fast decay mainly within 1 min, and another that shows a stable endurance for more than 20 min. The fast decaying component could be related to motor working memory that depends on factors like attention, while the second, more stable component could hold information acquired through procedural learning mechanisms. It is interesting to note that although in a typical prism adaptation/readaptation experiment the adaptation disappears rapidly after removing the prisms, the continuous training with prisms across several days or weeks result in the acquisition of a second or even more visuomotor correlations that can be accessed in the long term [9,10,21]. The consolidation of a second long-term visuomotor correlation suggests that after each complete adaptation/readaptation process some information remains, even if the subject had already went back to the original baseline visuomotor performance. We propose that the long-term component reported in the present experiment is different from the long-term formation of a second visuomotor calibration. One important difference would be that the long-term component we are describing tend to disappear as soon as the subject starts a visuomotor interaction without the prisms. In contrast, subjects that have acquired a longterm second visuomotor calibration can quickly switch between the old and the new calibrations depending on different factors like the presence or absence of the prisms frames [9]. However, it could be possible that some information stored during the second component described in this article could contribute to the acquisition of a long-term visuomotor correlation of the type obtained through extensive training across many sessions. It has been previously shown that under normal visuomotor reafference conditions complete aftereffect decay can be observed in less than 1 min of continuous interactions. Here we have shown that increasing the number of motor reafferences of the same nature as those used during training can accelerate the aftereffect decay in the absence of visual feedback. We also showed that under passive conditions it is possible to observe two different phases within the aftereffect. An initial period is characterized by a fast decay of about 40% of the total magnitude followed by a second period in which there is almost no extinction of the aftereffect.
J. Ferna´ndez-Ruiz et al. / Cognitive Brain Research 20 (2004) 92–97
Acknowledgements We thank Silvia Revuelta and Rafael Ojeda for their help in testing the subjects. This work was supported by CONACyT 34817-M, 30970-M and DGAPA IN210300.
[10] [11] [12]
References [1] J.S. Baizer, I. Kralj-Hans, M. Glickstein, Cerebellar lesions and prism adaptation in macaque monkeys, J. Neurophysiol. 81 (1999) 1960 – 1965. [2] C.S. Choe, R.B. Welch, Variables affecting the intermanual transfer and decay of prism adaptation, J. Exp. Psychol. 102 (1974) 1076 – 1084. [3] R. Dewar, Adaptation to displaced vision—the influence of distribution of practice on retention, Percept. Psychophys. 8 (1970) 33 – 34. [4] J. Fernandez-Ruiz, R. Diaz, Prism adaptation and aftereffect: specifying the properties of a procedural memory system, Learn. Mem. 6 (1999) 47 – 53. [5] J. Fernandez-Ruiz, C. Hall-Haro, R. Diaz, J. Mischner, P. Vergara, J.C. Lopez-Garcia, Learning motor synergies makes use of information on muscular load, Learn. Mem. 7 (2000) 193 – 198. [6] S. Kitazawa, T. Kimura, T. Uka, Prism adaptation of reaching movements: specificity for the velocity of reaching, J. Neurosci. 17 (1997) 1481 – 1492. [7] S.T. Klapp, S.A. Nordell, K.C. Hoekenga, C.B. Patton, Long-lasting aftereffect of brief prism exposure, Percept. Psychophys. 15 (1974) 399 – 400. [8] A.S. Kornheiser, Adaptation to laterally displaced vision: a review, Psychol. Bull. 83 (1976) 783 – 816. [9] T.A.Martin,J.G.Keating,H.P.Goodkin,A.J.Bastian,W.T.Thach,Thach, Thach, Throwing while looking through prisms: II. Specificity and
[13]
[14] [15]
[16]
[17] [18]
[19]
[20] [21]
[22]
97
storage of multiple gaze-throw calibrations, Brain 119 (Pt 4) (1996) 1199 – 1211. B.O. McGonigle, J. Flook, Long-term retention of single and multistate prismatic adaptation by humans, Nature 272 (1978) 364 – 366. G.M. Redding, Visual adaptation to tilt and displacement: same or different processes? Percept. Psychophys. 14 (1973) 193 – 200. G.M. Redding, Decay of visual adaptation to tilt and displacement, Percept. Psychophys. 17 (1975) 203 – 208. G.M. Redding, B. Wallace, Components of displacement adaptation in acquisition and decay as a function of hand and hall exposure, Percept. Psychophys. 20 (1976) 453 – 459. G.M. Redding, B. Wallace, Adaptive Spatial Alignment, vol. x, Lawrence Erlbaum Associates, Mahwah, NJ, 1997 194 pp. Y.K.K. Rossetti, T. Mano, Prismatic displacement of vision induces transient changes in the timing of eye – hand coordination, Percept. Psychophys. 54 (1993) 355 – 364. L. Squire, Declarative and nondeclarative memory: multiple brain systems supporting learning and memory, in: P. Andersen (Ed.), Memory Concepts. Basic and Clinical Aspects, Excerpta Medica, New York, 1993, pp. 3 – 25. E. Taub, L.A. Goldberg, Prism adaptation: control of intermanual transfer by distribution of practice, Science 180 (1973) 755 – 757. E. Tulving, Human memory, in: P. Andersen (Ed.), Memory Concepts. Basic and Clinical Aspects, Excerpta Medica, New York, 1993, pp. 27 – 45. M.J. Weiner, M. Hallett, H.H. Funkenstein, Adaptation to lateral displacement of vision in patients with lesions of the central nervous system, Neurology 33 (1983) 766 – 772. R.B. Welch, Perceptual Modification: Adapting to Altered Sensory Environments, vol. xv, Academic Press, New York, 1978 346 pp. R.B. Welch, B. Bridgeman, S. Anand, K.E. Browman, Alternating prism exposure causes dual adaptation and generalization to a novel displacement, Percept. Psychophys. 54 (1993) 195 – 204. C.R. Hamilton, J. Bossom, Decay of prism aftereffects, J. Exp. Psychol. 67 (1964) 148 – 150.