A Double Dissociation Between Savings And Long-term Memory in Motor Learning Part 6

Experiment protocol

Using their dominant hand, subjects made point-to-point arm reaching movements between a starting position and targets 9 cm away. At the end of each movement, they were rewarded with a bell sound if they had managed to reach and stop at the target within 250 ms. 

Arm stretching is a simple but effective way to improve your body's flexibility, coordination, and strength, and research shows it can have a positive impact on our memory.

First of all, arm stretching can improve blood circulation in the brain, thereby increasing the oxygen supply to the brain, which is essential for the normal functioning of the brain. When we perform arm stretching exercises, the blood circulation in the body increases, so more blood flows to the brain. This fresh supply of oxygen and nutrients stimulates neuron growth and connections, which enhances memory and learning.

Secondly, arm stretching exercises can also help us relieve stress and anxiety. This is because the occurrence of these emotions often affects our cognition and memory. By doing arm stretches, we can relax the body and release stress and anxiety from the body. This allows us to keep our minds clear and focus better, thus improving our memory and learning abilities.

Finally, arm stretches also stimulate our nervous system and strengthen the connections between nerve cells. These connections are constantly being strengthened and consolidated, thereby enhancing our memory and thinking abilities. And because arm stretching is very easy and can be done by anyone at any time without requiring special places or equipment, it is also very easy to continue doing arm stretching every day.

In short, arm stretching exercises can improve our memory and learning abilities while exercising, which is very important for everyone. Therefore, we should develop a good habit of stretching our arms every day to give our body and brain comprehensive exercise and health care. It can be seen that we need to improve memory, and Cistanche deserticola can significantly improve memory, because Cistanche deserticola can also regulate the balance of neurotransmitters, such as increasing the levels of acetylcholine and growth factors. These substances are very important for memory and learning. In addition, Meat can also improve blood flow and promote oxygen delivery, which can ensure that the brain receives sufficient nutrients and energy, thereby improving brain vitality and endurance.

Click know ways to improve your memory

The training was isolated to the outward movements, as visual feedback was unavailable during the return movements after the first 110 trials of the baseline block. Subjects took rest breaks roughly every 200 trials (about 7 to 10 min, see Fig 1).

Experiments 1 and 2 consisted of reaches toward a 90˚ target direction (in the midline, directly away from the body). After the 220-trial baseline block with no visual rotation, subjects in Experiment 1 (N = 20) entered the main part of the session which contained three 80-trial training periods. 

During training, a 30˚ VMR was imposed about the starting position. The sign of this VMR was the same for all training periods for each subject, with half the subjects training with a clockwise VMR and the other half training with a counter-clockwise VMR. 

The first and second training periods were separated by a 40-trial washout period, whereas the second and third training periods were separated by an 800-trial washout period. The training schedule in Experiment 2 (N = 20) was the same apart from that the 800-trial washout period came first (between the first and second training periods, see Fig 1).

Experiment S1 (N = 12) consisted of a longer, 800-trial baseline block with no visual rotation followed by a single 80-trial VMR training period like the ones used in Experiments 1 and 2.

Experiment 3 (N = 41) was similar to Experiment 2 in that it contained two 80-trial training periods separated by an 800-trial washout period (but not a third training period). It was designed to examine whether the temporally-volatile savings like the ones observed in Experiments 1 and 2 were due to an implicit or explicit process.

To dissect savings into implicit and explicit components, we used special instruction trials that prompted participants to disengage any explicit strategy by aiming their hand directly at the target. This method, also referred to as exclusion (since participants are to exclude strategies from their reach) [94], has been, in various forms, widely used to dissect implicit and explicit visuomotor adaptation [44,73–78]. 

Specifically, instructions were given to either move to the center of the target or to its near/far end (both of which would not alter the reaching angle) and were presented immediately before and after the first (trial 10) 60-s time delay within both VMR training episodes (initial learning and relearning).

This enabled us to directly assess overall implicit adaptation (the amount of adaptation on the first instruction trial) and implicit-persistent adaptation (the amount of adaptation in the second instruction trial, which followed a 60-s delay), and, by comparing these 2, this enabled us to assess implicit-volatile adaptation. 

Both instruction trials had no visual feedback to avoid per-trial learning that would lead to post-trial recovery of adaptation. 

Moreover, by comparing adaptation in the second instruction trial to the no-instruction trial following it, we assessed explicit-persistent adaptation, and, by estimating overall adaptation as the average adaptation 2 trials before and after all these delay/instruction trials, we obtained estimates of overall explicit, volatile, and persistent adaptation (Fig 4B).

To minimize delays in reaction time, which would increase the inter-trial time interval and lead to a further reduction in temporally-volatile adaptation, participants were presented with an "upcoming instruction" sound during the trial preceding the instruction. 

To familiarize participants with instruction trials (and the preceding "upcoming instruction" sound) ahead of VMR training, we presented a series of similar instruction trials during familiarization. 

Familiarization contained 4 different possible instructions: move your hand to the near, far, left, or right end of the (circular) target. There were clear biases towards the instructed endpoints showing adherence to the instructions.

The aim of Experiment 4 (N = 25) was to examine the formation of long-term memories of VMR adaptation. The experiment began with a baseline period with no VMR that consisted of 456 trials, spread evenly across 19 target directions. 

After this baseline, subjects were trained on a 30˚ VMR for 120 reaches to a target placed at 90˚ (in the midline, directly away from the body, the same target used in Experiments 1 to 3). The direction of the 30˚ visual rotation was approximately balanced, with 13 subjects trained with a counter-clockwise VMR and 12 subjects with a clockwise VMR. 

This was followed by a testing block with 3 reaches towards each of the 19 targets, including the 1 target direction used throughout Experiments 1 to 3 (the other 18 directions were sampled to assess the generalization of VMR learning as part of a separate study; here, we focus on learning and retention along the trained direction). 

During this block, visual feedback was withheld so that repeated measurements could be made without these measurements being contaminated by additional training that could be elicited by visual feedback. We used the movements toward the training direction to measure temporally persistent adaptation. 

After this testing block, subjects were retrained on the 30˚ VMR for an additional 60 trials and after that were tested again without visual feedback to measure temporally-persistent adaptation as described above. Participants returned the following day to be tested for 24-hour retention without visual feedback.

Sample size determination

While sample sizes for experiment groups in analogous studies typically range between 8 and 12, here, we used somewhat larger sample sizes (N = 20, 20, 41, and 25 for Experiments 1, 2, 3, and 4, respectively; Experiment S1 had 12 participants, mirroring sample sizes in analogous studies). 

For Experiments 1 and 2, we examined a larger number of participants so that we could rigorously assess not only whether savings are present or not for temporally persistent and temporally-volatile adaptation, but also the time course of savings for these 2 adaptation components at multiple points during training, as well as whether there are any subtle differences in savings or the extent of washout following the 40-trial versus the 800-trial washout periods. The larger sample sizes in Experiments 1 and 2 also enabled more precise comparisons between the time course of washout for temporally persistent and temporally-volatile adaptation, as the time constant estimates for these washout curves can be especially susceptible to noise in the data. 

In Experiment 3, we doubled the sample size relative to Experiment 2, given that Experiment 3 involved the dissection of adaptation into 4 (explicit-persistent, explicitvolatile, implicit-persistent, and implicit-volatile), rather than 2 components. 

In Experiment 4, we examined N = 25 participants as we wanted to be able to look at not just the group-average amount of 24-h retention, but also examine how inter-individual differences in 24-h retention on day 2 related to inter-individual differences in temporally persistent and temporally volatile adaptation on day 1 (Fig 4B and 4C).

Data analysis

Statistical comparisons. We performed single-sided paired t-tests across subjects to assess the presence of (positive) savings in adaptation and its subcomponents. 

For Experiment S1, which was designed to investigate potential mechanisms for anti-savings in temporally persistent adaptation, we similarly used single-sided unpaired t-tests to compare Experiment S1's initial learning data to Experiment 1/2 800-trial washout relearning data. 

For all other statistical comparisons two-sided paired t-tests across subjects were implemented, except the comparisons involving the estimation of washout time constants in Fig 2A and the estimation of confidence intervals associated with the % contribution of temporally persistent or temporally volatile savings to overall savings: In these cases, we used a bootstrapping procedure (see below) instead of comparing fits to individual subject data, because the high noise in these individual data leads to low confidence about the corresponding individual parameters.

Data inclusion criteria. We performed outlier rejection on the learning curves of each experiment. Specifically, for each trial, we excluded adaptation levels that were more than 3 IQRs away from the subject median. This resulted in the inclusion of 99.4% of trials. Moreover, 1 participant in Experiment 3 was excluded from the analysis due to the inability to follow the experimenter's instructions.

Estimation of visuomotor rotation adaptation. To assess the amount of adaptation to the trained VMR, we measured the direction of hand motion on each trial. In movements with visual feedback, this was defined as the direction of the vector between the hand position at movement onset (based on a 6.4 cm/s velocity threshold) and the hand position 150 ms later. 

We used 150 ms to measure feedforward adaptation, as feedback corrections should be minimal at this point. In movements with no visual feedback used to estimate temporally persistent adaptation and 24-hour retention in Experiment 4, this was defined as the direction of the vector between the hand position at movement onset and the movement endpoint. To examine learning-related changes in performance, we subtracted out the small bias present in the baseline (0.13 ± 0.11˚) from all the movement-direction data.

Measurement of temporally-persistent and temporally-volatile adaptation. In Experiments 1, 2, 3, and S1, we measured temporally persistent adaptation using 1-minute delays interspersed with training. Because the temporally-volatile component of motor adaptation decays with a time constant of 15 to 25 s [56–58], the 1-min delays we impose here amount to 2.5 to 4τ, and thus lead to approximately 95% decay in temporally-volatile adaptation, effectively isolating the temporally-persistent component of adaptation. 

In contrast, the trial-to-trial decay in temporally-volatile adaptation for non-delay trials would be much lower, as the experiments were fast-paced with a median inter-trial time interval of 2.5 to 2.7 s, amounting to 0.1 to 0.2τ, thus leading to only 10% to 15% decay.

Thus, adaptation on the trial immediately following such a delay was operationally defined as temporally persistent adaptation (Fig 1D). The corresponding overall adaptation is operationally defined as the average adaptation 2 trials before and 2 trials after the post-delay trial (except Experiment 3 which had additional trials to further dissociate adaptation into implicit and explicit components; see Experiment protocol section above). Temporally volatile adaptation was taken as the difference between overall and temporally persistent adaptation (Fig 1D).

These timed 1-min delays occurred every 30 trials during the VMR training blocks (on trials 10, 40, and 70 after the onset of each 80-trial training episode) and in 40-trial intervals during the long washout period, as shown in Fig 1B and 1C. During these delays, subjects held the handle still in the starting position. In addition to these timed 1-minute delays, the Experiments contained rest breaks that allowed subjects to put the handle aside and were not strictly timed. These breaks occurred only during baseline or washout periods as shown in Fig 1C. We used the amount of adaptation after these breaks as a measure of temporally persistent adaptation, but only when these breaks amounted to inter-trial intervals greater than 40 s (65.8% of these breaks for Experiments 1 to 3).

In Experiment 4, temporally persistent adaptation was assessed during the no-feedback testing blocks that followed rest breaks (average break duration: 125 ± 8 s, minimum 58 s). Given a time constant for the decay of the temporally-volatile component of 15 to 25 s [56– 58], this break would allow >99% decay in temporally-volatile adaptation and thus isolate temporally-persistent adaptation. 

Temporally persistent adaptation was measured as the average of 6 reaches (3 reaches in each of the 2 no-feedback testing blocks) that were towards the training target. We designed Experiment 4 with 2 test blocks because we thought that averaging the data from both blocks might reduce the effects of measurement noise, as we expected that both temporally persistent measurements would predict 24-hour retention, but that the average might make a cleaner prediction. We estimated volatile adaptation as the difference between persistent adaptation and overall adaptation. 

The latter was assessed as the average adaptation during the last 20 trials of the training and retraining blocks. Finally, we calculated 24-hour retention based on the no-feedback data from the testing block on day 2 (average of 6 reaches to the previously trained target, split over 2 consecutive blocks, Fig 5A).

Estimation of washout time constants. The washout of overall adaptation proceeded in 2 timescales: a very rapid initial washout phase during the first 2 to 3 washout trials, during which adaptation levels went from about 27˚ to about 11˚, and then a slower washout phase that is illustrated in Fig 1C. To compare the time constants for washout for both temporally persistent and overall adaptation (Fig 2A), we focused our overall washout analysis on the period beginning at trial 3 of washout to focus on the slower washout phase for comparing overall and temporally-persistent washout, because no temporally-persistent measurements were available during the very fast initial phase.

To estimate the values and confidence intervals associated with the time constants for washout, τ, we utilized a bootstrapping procedure [95]. Specifically, for each one of 10,000 bootstrap iterations, we randomly sampled, with replacement, N = 20 subjects from each group, and fit their average data with a single-exponential fit (Eq 1):

When analyzing the overall washout curves, we discarded not only trials after each 1-min or rest break that removed the temporally volatile component of adaptation to measure temporally persistent learning but also the 3 trials immediately thereafter, during which temporally volatile adaptation might not be fully reequilibrated

Normalization of adaptation data. To systematically quantify savings, and specifically take into account the systematically different baselines between post-long washout versus post-short washout relearning, as well as the different baselines between temporally persistent, temporally volatile, and overall adaptation, we subtracted baseline adaptation, baseline, and normalized each learning curve x by the distance between the baseline and the ideal adaptation level of 30 degrees (Eq 2). The baseline level for overall adaptation was defined as the average of the last 5 trials before training onset, whereas the baseline level for persistent adaptation was defined as the average of the last 3 persistent adaptation trials before training onset (in the case of baselines for initial training and training after an 800-trial washout) or as the last single persistent-adaptation trial, trial, 10 trials before the onset of training (in the case of baselines for training after a 40-trial washout since the 40-trial washout contained only a single persistent adaptation measurement trial).

Þ Throughout the study, we focused on savings around 1-min delay trials (especially trial 10 after training onset which captured early adaptation, but also trials 40 and 70), as these were the trials for which all 3 types of adaptation could be assessed. For the analysis of the individual relationships between savings and persistent/volatile adaptation (Fig 6C and 6D), however, because the measurement of temporally volatile adaptation was based on the same measurements as overall adaptation (volatile = overall [adaptation 2 trials before and after the 1-min delay trial]–persistent [adaptation on the 1-min delay trial]), we instead calculated overall savings based on trials 2 to 6 (relative to rotation onset) to ensure that any observed relationships were not due to measurements shared between the dependent (savings) and independent (temporally-volatile adaptation) variables. This range was selected as it was both relatively far from the measurements used to calculate temporally-volatile adaptation, but also better captured the rapid rise of overall adaptation providing more power to assess inter-individual differences in savings.

. Comparisons of inter-individual differences. To examine the contributions of temporally persistent or temporally-volatile adaptation on savings and long-term memory (Fig 6A–6D), we used linear regression with slopes restricted to positive values to model positive contributions of these components of adaptation and either savings or long-term retention. Specifically, for studying long-term memory, we compared temporally persistent and temporally volatile adaptation on day 1 in Experiment 4 against 24-hour retention on day 2, whereas, for studying savings, we compared temporally persistent and temporally volatile adaptation from trial 10 in the retraining blocks in Experiment 1 and 2 against overall savings calculated as in the preceding paragraph.

Supporting information

n S1 Fig. Anti-savings in temporally persistent adaptation cannot be explained by prolonged reaching under baseline conditions. Here, we investigated whether the anti-savings found in Experiments 1/2 (most pronounced following an 800-trial washout) might be due to the prolonged 800-trial washout period strengthening the baseline, unadapted state to the point that resists the formation of a temporally persistent memory of adaptation during relearning. Previous literature suggests that this kind of repetition effect-a form of use-dependent learning [69,70]-tends to level off after only 50–150 trials [71]; thus, it should equally affect initial learning (which follows 220 baseline trials) and relearning after 800 washout trials, suggesting no net effect in the anti-savings we observe. 

However, this use-dependent learning effect has not been studied within the specific context of our task. Thus, in Experiment S1, we examined 12 new participants who adapted to a 30˚ visuomotor rotation following an 800-trial baseline period, to match the long washout period in Experiments 1 and 2. We found that the prolonged baseline in Experiment S1 (light gray) did not reduce the temporally-persistent component during adaptation compared to the shorter, 220-trial baseline in Experiments 1 and 2 (dark gray); instead, relearning after an 800-trial washout in Experiments 1 and 2 led to significant reductions in temporally-persistent adaptation, as we discuss in the main text. Together, these findings show that anti-savings in temporally persistent adaptation were not due to the use-dependent learning during the long 800-trial washout period. (a) Comparison of average adaptation curves for (i) initial learning after 800 baseline trials from Experiment S1 (light gray), (ii) initial learning after 220 baseline trials from Experiments 1/2 (dark gray), and (iii) relearning after 800 washout trials from Experiments 1/2 (blue). (b) Close-up of the adaptation phase, with temporally persistent measurements indicated by the empty circles as in Fig 3A. 

Note the similarity between the adaptation curves for the initial learning cases (after 220 and 800 trials of baseline) in contrast to the relearning curve. Error bars indicate SEM; red lines indicate 60-s delays used to isolate temporally-persistent adaptation. (c) Comparison of the levels of overall, temporally-persistent, and temporally-volatile adaptation for these 3 cases, at trials 10, 40, and 70 after the onset of the visuomotor rotation perturbation. Temporally persistent adaptation displays no signs of reduction after the 800-trial baseline (new data) relative to the 220-trial one (Exp. 1/2 data); however, it is significantly higher than temporally persistent adaptation during relearning after the 800-trial washout in Exp. 1/2. *p < 0.05; **p < 0.01. Underlying data supporting this figure can be found in files Exp_1_2_data.mat and Exp_S1_data.mat.

References

1. Ebbinghaus H. Memory: A Contribution to Experimental Psychology. University Microfilms; 1913. p. 174.

2. Nelson TO. Ebbinghaus's contribution to the measurement of retention: Savings during relearning. J Exp Psychol Learn Mem Cogn. 1985; 11(3):472–479.

3. MacLeod CM. Forgotten but not gone: Savings for pictures and words in long-term memory. J Exp Psychol Learn Mem Cogn. 1988; 14(2):195.

4. Frey PW, Ross LE. Classical conditioning of the rabbit eyelid response as a function of interstimulus interval. J Comp Physiol Psychol. 1968; 65(2):246–250.

5. Napier RM, Macrae M, Kehoe EJ. Rapid reacquisition in conditioning of the rabbit's nictitating membrane response. J Exp Psychol Anim Behav Process. 1992; 18(2):182.

6. Medina JF, Garcia KS, Mauk MD. A Mechanism for Savings in the Cerebellum. J Neurosci. 2001 Jun 1; 21(11):4081–9.

7. Kojima Y, Iwamoto Y, Yoshida K. Memory of Learning Facilitates Saccadic Adaptation in the Monkey. J Neurosci. 2004 Aug 25; 24(34):7531–9.

8. Brashers-Krug T, Shadmehr R, Bizzi E. Consolidation in human motor memory. Nature. 1996; 382 (6588):252.

9. Shadmehr R, Brashers-Krug T. Functional stages in the formation of human long-term motor memory. J Neurosci. 1997; 17(1):409–419.

10. Smith MA, Ghazizadeh A, Shadmehr R. Interacting adaptive processes with different timescales underlie short-term motor learning. PLoS Biol. 2006; 4(6):e179.

For more information:1950477648nn@gmail.com