187
NeuroRehabilitation 36 (2015) 187–194 DOI:10.3233/NRE-151206 IOS Press
A comparison of self-generated versus taught internal strategies for working memory Diane Powers Dirette∗ Western Michigan University, 1903 W. Michigan Ave., Kalamazoo, MI 49008, USA Tel.: +1 269 387 7243; Fax: +1 269 387 7262; E-mail: [email protected].
Abstract. BACKGROUND: Internal strategies are effective for improving working memory. These internal working memory strategies can be taught or self-generated. OBJECTIVE: This study compares working memory performance using taught versus self-generated internal working memory strategies and explores the quantity, type and carry-over of the use of these strategies. METHODS: An experimental cohort design with randomly assigned groups compared the performances among 120 participants on 5 memory tests given prior, immediately following and at 1 month post intervention. RESULTS: There were no significant differences in the number or type of internal memory strategies used by the groups with chunking, repetition and association used most commonly. The group that self-generated strategies performed significantly better on a contextual memory test and the group that was taught strategies performed better on a face/name recall test for which a specific strategy was taught. The group that was taught strategies performed significantly worse on contextual memory tests from pre-test to follow-up. CONCLUSIONS: Participants who generated their own strategies or used a specific strategy for a specific task did as well as participants who were taught myriad internal WM strategies. Teaching too many strategies may overwhelm participants or may distract them from using the context that is available in such tasks. Keywords: Cognitive rehabilitation, mnemonics, short-term memory
1. Introduction Working memory (WM) is “the ability to hold information in mind and to manipulate it in light of incoming material” (p. 1450, McAllister, et al., 2006). Because the neural substrates of WM are in regions commonly damaged in traumatic brain injuries (TBI), deficits in WM are common following even mild TBIs (BarkerCollo & Feigin, 2008; McAllister, et al., 2006). Deficits in WM are also common with some progressive neurological disorders such as Mutiple Sclerosis (Dirette, 2012) and Parkinson’s Disease (Naismith, Mowszowski, Diamond & Lewis, 2013). In addition to being a common deficit, WM is resistant to improvement or healing over time, even with intensive restorative therapy (Lippert-Gruner, Kuchta, Hellmich & Klug, 2006). Deficits in WM significantly impact activities
of daily living such as employment, housework, and meal preparation (Dombovy & Olek, 1996), and nearly every aspect of learning (Anderson, Catroppa, Morse & Haritou, 1999). Cognitive rehabilitation is a collaborative multidisciplinary service that often focuses on the improvement of WM. The traditional cognitive rehabilitation paradigm uses restoration and compensatory strategy training methods to improve WM (Ylvisaker, Hanks & Johnson-Greene, 2002). Restorative interventions are typically used first and, if unsuccessful, followed by training in the use of compensatory strategies (Yivisaker, et al., 2002). Compensatory strategies can be either internal or external. Internal WM strategies are techniques used within one’s own mind to enhance the processing and recall of information.
1053-8135/15/$35.00 © 2015 – IOS Press and the authors. All rights reserved
188
D.P. Dirette / A comparison of self-generated versus taught internal strategies for working memory
Some examples of internal memory strategies include chunking, pacing, verbalization, association, acrostics, method of loci, pegword, rehearsal and imagery. Chunking is mentally grouping information into segments to condense the information (Diller, 1993). Pacing is performing a task with intermittent pauses to decrease the effects of proactive interference (Anderson, 1985; Toglia, 1993). Verbalization is orally repeating information gathered from visual sources (Toglia, 1993). Association is creating relationships between the information that is presented and previously learned information. Acrostics are creating expressions using the first letter of each word that needs to be remembered. Method of loci is using association with a known physical layout such as one’s home or office. Pegword is creating a list of words associated with sequential numbers and then using association with the words that were created (e.g. one-fun-shoe with the association that you may love to kick off your shoes and have fun). Rehearsal is mentally repeating the information that was presented. And, imagery is creating a mental picture of information that is presented orally. The term “mnemonics” is often used in the literature to describe internal WM strategies. Mnemonics are any techniques used to help remember information. Short rhymes or phrases that elaborate on the information are often thought of when the term mnemonics is used. Repetition or rehearsal of information is not generally considered a mnemonic as there is an implication in mnemonics that something about the information is elaborated on to help with encoding and recall. 1.1. Literature review Although cognitive rehabilitation has been shown to be effective for improving WM through the use of compensatory strategies (Cicerone, Dahlberg, Kalmer, et al., 2000; Cicerone, Dahlberg, Malec et al., 2005), the most effective means for developing the use of these strategies has not been established (Rasmussen, Treit, & Pei, 2010). Internal memory strategies can be taught by therapists or people may develop their own strategies. The development of strategies without instruction is referred to as self-generation. Cognitive rehabilitation therapists typically teach people myriad strategies and allow them to practice and choose the ones that work the best for them. According to the generation effect, however, optimal acquisition and retention of information is best established though active participation rather than by passive observation (Schefft, Dulay, & Fargo, 2008). It is unknown if this effect applies to
self-generation of internal strategies or if the traditional method of teaching strategies is more effective. Studies typically compare an intervention that teaches compensatory strategies to an intervention that is deemed restorative or remedial. In the remedial intervention, the participants are asked to repeatedly perform a cognitive task, but they are not given instruction in the use of strategies to complete the task. For example, a randomized, controlled trial of 61 adults who had Multiple Sclerosis with mild cognitive deficits, compared an intervention focused on training in the use of strategies to a remedial intervention (Stuifbergen, Becker, Perez, Morrison, Kullberg, & Todd, 2012). Participants were tested initially and at the end of treatment using The Minimal Assessment of Cognitive Function. Both groups improved significantly and there were no significant differences in performance except that the strategy group did significantly better on a test of verbal memory. The use of self-generated strategies by the remedial group, however, may have confounded the results as there was no measure of strategy use by the remedial group. A couple of studies have compared WM performance for taught versus self-generated strategy groups. Derwinger, Stigsdotter Neely and Backman (2005) compared the memory performance of 28 older adults (61–83 years old) who were taught a mnemonic for recall of a 4 digit number to that of 29 age-matched participants who were encouraged to use self-generated strategies to recall the same numbers. The study also included a control group of 24 participants who received no training. All three groups improved equally on memory tests during the study. In the long-term follow-up testing, however, the performance of the mnemonic group decreased, the performance of the self-generated group improved, and the performance of the control group stayed the same. The researchers concluded that self-generated strategies may be more beneficial to older adults than a taught strategy because selfgenerated strategies do not require practice to maintain their use. Wang and Thomas (2000) explored the differences in serial word recall among 202 adult college students who were divided into three groups, two mnemonic groups and one control group. The two mnemonic groups included one group who was taught the method of loci and one group who was taught the pegword system. The control group was told to use their own best strategies. After a two-day delay, the three groups demonstrated no significant differences in their ability to recall the serial word lists.
D.P. Dirette / A comparison of self-generated versus taught internal strategies for working memory
These studies provide preliminary exploration of the differences in self-generated versus taught internal strategies for WM, but they offer limited exploration of the carry-over of the use of the strategies, the types of strategies that were used by the groups or the types of tasks that may be enhanced by the strategies that are used. The participants who were taught strategies may have added self-generated strategies to the process, as well. The purpose of this current study is to compare the immediate and long-term performance that results from taught versus self-generated internal WM strategies and to compare the quantity and type of strategies that are used by the two groups.
2. Method 2.1. Design An experimental cohort design was used to examine the differences in internal WM strategy use with groups randomly assigned to be either control or intervention groups. Data were not collected simultaneously. Participants were blind to their group assignment and the test scoring was completed by a research assistant who was blind to group participation. 2.2. Participants All participants were students at the university at which the study was conducted. HSIRB approval was obtained for the recruitment and study procedures. Participants were recruited at the beginning of a classroom session without the instructor present and given the option to participate in the study or leave the classroom. The instructors were never given any information about attendance or participation. All participants provided individual informed consent and completed a demographic data sheet. The testing and intervention sessions were completed in a group format. 2.3. Procedures Pre-tests were given at the beginning of each session for all groups before any interventions were provided. Immediately following the pre-tests, the intervention group participated in a 30 minute training session on the use of internal WM strategies. The control group participated in the same length of intervention time by practicing the same activities that were used with the intervention group. The control group, however,
189
received no training in the use of WM strategies during the activities. The internal WM strategies that were taught to the intervention group during this study included chunking, pacing, verbalization, association, acrostics, method of loci, peg word, rehearsal and imagery. These 9 strategies were taught in the order listed. Each strategy was taught using an activity. Chunking was taught using a series of numbers such as those for a phone number or credit card, pacing was taught using a short story, verbalization was taught using addresses and phone numbers, association was taught using names of people and places and acrostics was taught using a series of words such as the Great Lakes, the cranial nerves and the planets. Post-tests were given immediately after the intervention and repeated at follow-up, one month following the intervention. Different versions of each test were given at each testing session to reduce the risk of a practice effect. A standardized order and script were used during all testing sessions. The participants were not allowed to write down any information, use electronic devices or make any notations during any of the assessments. In other words, the use of external WM strategies was not allowed. At the end of the follow-up session, the participants completed a brief questionnaire about their use of internal memory strategies during the testing procedures. They were asked to identify the number and types of strategies that were used. 2.4. Instruments Five memory tests were given to each group at pre-test, post-test and follow-up. The five tests included Face/Name Recall, the Contextual Memory Test (CMT), the Grey Oral Reading Test, 4th Edition (GORT-4), the California Verbal Learning Test (CVLT II) and the Rey-Osterrieth Complex Figure Test (ROCF). Face/Name Recall was given using two faces/names at pre-test, and four faces/names at post-test and followup. The pictures, names and procedures that were used were similar to those used in the Rivermead Behavioural Memory Test (RBMT) with a black and white picture of a face shown for five seconds on a document camera and the first and last name spoken at the beginning of that visual presentation. The participants were then asked to recall the names of the people when shown the faces after a 20 minute delay. The RBMT has been shown to have good reliability and validity
190
D.P. Dirette / A comparison of self-generated versus taught internal strategies for working memory
with varied populations (DeWall, Wilson, & Braddeley, 1994; Kucukdeveci, Kutlay, Elhan, & Tennant, 2008; Yassuda, et al., 2010). The CMT is a standardized test of visual memory that uses a picture card of 20 related, everyday objects. The picture card is presented for 90 seconds and then the participant is asked to recall as many objects as possible. The participants in this study were asked to write these objects on a list instead of recalling them verbally. Test-Retest reliability scores for the CMT range from 0.74 to 0.94 (Toglia, 1993). The concurrent validity with the RBMT screening scores range from 0.80 to 0.85 (Toglia, 1993). Paragraphs 8, 9 and 10 from the GORT-4 (Wiederholt & Bryant, 2001) were used for pre-tests, post-tests and follow-up tests respectively to assess recall of written information. The participants were given time to read the paragraphs to the end and then asked to recall as much information as possible. Recall was scored using the measurement of information bits written down by the participant immediately following reading the paragraph. Order of the story recall was not considered. The reliability and validity measures for the GORT-4 do not apply to this study as the test was not used with the test questions, but used as a measure of information recall as it is done in the RBMT. The CVLT II was administered as a measure of auditory recall. The list of 16 words was read to the participants at an interval of one word per second. The participants were asked to immediately recall and write down as many items from the list as they could remember. As told to the participants, order was not considered. The same list was then repeated in the same order and at the same rate. After the list was read for the second time, the participants were asked to write down on a new piece of paper as many items as they could remember the second time. The test/re-test reliability scores for the CVLT II range from 0.80–0.84 for the same version of the test and 0.61–0.73 for an alternative version of the test (Woods, Delis, Dramer, & Holdnack, 2006). Internal consistency for the CVLT II ranges from.78 to.94 (Delis, Kramer, Kaplan, & Ober, 2000). The ROCF is an assessment of visuospatial recall in which participants are asked to reproduce a complicated line drawing. The ROCF was scored using the formal scoring criteria for the standard 18 elements with 0.5–2.0 points awarded for each reproduced element depending on accuracy/distortion and location of the reproduction. The ROCF has been shown to have good validity (Smith, et al., 2007) and inter-rater reli-
ability scores ranging from 0.92–0.97 (Loring, Martin, Meador, & Lee, 1990; Luzzi, et al., 2001). Figure A was used for pre-tests and follow-up tests and Figure B was used for post-tests. 2.5. Data analyses All data were entered into SPSS version 17 for the analysis. Variables were analyzed for normal distributions. Where skewed variables occurred, a non-parametric analysis was used to make group comparisons. All other group comparisons were completed using t-tests or RMANOVA as appropriate. In cases where tests of homogeneity were not met, the equality not assumed or the Greenhouse-Giesser correction was used. Alpha was set at 0.05 as is appropriate for comparing groups of this size.
3. Results 3.1. Demographics Altogether there were 120 participants in the study with a mean age of 22.24 years (sd = 5.61) with a range of 18–58 years. There were 50 participants in the control group and 70 participants in the intervention group. The mean age of the control group was 23.81 years (sd = 6.52) with a range of 19–58 years. The mean age of the intervention group was 21.19 years (sd = 4.66) with a range of 18–50 years. The age of the control group was significantly higher (U = 967.5, sign. = 0.000). All of the participants were female except for four male participants in the control group and five male participants in the intervention group. The assessments and interventions were all completed in three control and three intervention groups. The size of the groups ranged from 11 to 28 participants. The sizes of the control groups were 11, 16 and 23 participants and the sizes of the intervention groups were 18, 24 and 28 participants. 3.2. Strategy use There was no significant difference (t = 0.886; sign = 0.378) between the total number of internal WM strategies used by the control group (X = 2.76; sd = 0.990) and by the intervention group (X = 2.59; sd = 0.960). A series of chi square analyses also revealed no significant differences in the types of internal WM strategies that were used by the groups. Chunking was the strategy used most often by both groups with 83%
D.P. Dirette / A comparison of self-generated versus taught internal strategies for working memory
191
of the intervention group and 80% of the control group reportedly employing this strategy. The second most commonly used strategy was repetition with 59% of the intervention group and 64% of the control group using this strategy. The third most common was association (39% of the intervention group and 46% of the control group), the forth was visualization (19% of the intervention group and 30% of the control group), the fifth was acrostics (17% of the intervention group and 20% of the control group) and the sixth was verbalization (14% of the intervention group and 4% of the control group). The seventh through tenth most commonly used strategies were method of loci, motor practice, pacing, and pegging with fewer than 10% of participants using those strategies in either group. One category of strategy use was deemed “other”. There was no significant difference in this category with 10% of the intervention group and 20% of the control group reportedly using a strategy that was not captured in the listed categories. The most common of the “other” strategies included sequencing and counting.
performance was significantly worse on these two WM measures. See Table 1. There was one significant difference between group scores at pre-test. The control group did significantly better on the CVLT II trial 1 (U = 1276.5, p = 0.011) and trial 2 (U = 1278.5, p = 0.011). The mean scores for the intervention group were trial 1 x = 7.09 (sd = 1.99) and trial 2 x = 11.33 (sd = 2.09). The mean scores for the control group were trial 1 x = 7.92 (sd = 1.82) and trial 2 x = 12.22 (sd = 2.21). See Table 2. There were no significant differences between group scores at post-test. See Table 3. There were only two significant differences between the groups at the one month follow-up. The control group did significantly better on the follow-up test of the CMT (U = 1255.0, p = 0.016) and the intervention group did significantly better on the follow-up text of Face/Name Recall (U = 907.0, p = 0.000). See Table 4.
3.3. Comparisons of scores
There were no differences in the number or type of internal WM strategies used by the participants who were taught (intervention group) and the participants who were given a remedial intervention during which they self-generated internal WM strategies (control group). Participants from both groups used between 2 and 3 different types of strategies during the follow-up testing. The control group participants self-generated and used the same strategies that were used by the inter-
Both groups improved significantly on the CVLT (trials 1 and 2) and the RCFT from pre-test to follow-up testing. The intervention group also improved significantly on Face/Name Recall, but the control group’s performance did not change significantly. The control group’s performance did not change significantly on the CMT or the GORT-4, but the intervention group’s
4. Discussion
Table 1 Comparisons of pre-test to follow-up scores Assessment
Intervention group t = 6.023 (p = 0.000)∗ t = 3.644 (p = 0.001)∗∗ t = 3.024 (p = 0.004)∗∗ t = 9.120 (p = 0.000)∗ t = 7.430 (p = 0.000)∗ t = 3.643 (p = 0.001)∗
Face/Name recall CMT GORT-4 CVLT II 1st trial CVLT II 2nd trial RCFT ∗ Significantly
Control group t = 1.512 (p = 0.137) t = 0.339 (p = 0.691) t = 2.001 (p = 0.051) t = 4.107 (p = 0.000)∗ t = 3.848 (p = 0.000)∗ t = 3.919 (p = 0.000)∗
better. ∗∗ Significantly worse. Table 2 Mean comparisons of pre-test assessments
Assessment Face/Name recall CMT GORT-4 CVLT II 1st trial∗ CVLT II 2nd trial∗ RCFT ∗ Significant
difference.
Intervention group
Control group
Comparison/value
1.55 (sd = 0.654) 14.69 (sd = 2.56) 14.37 (sd = 5.27) 7.09 (sd = 1.99) 11.33 (sd = 2.09) 35.87 (sd = 7.72)
1.66 (sd = 0.626) 14.72 (sd = 2.17) 16.06 (sd = 4.91) 7.92 (sd = 1.82) 12.22 (sd = 2.21) 35.36 (sd = 6.74)
U = 1559.5 (p = 0.277) U = 1749.0 (p = 0.996) U = 1441.0 (p = 0.099) U = 1276.5 (p = 0.011) U = 1278.5 (p = 0.011) U = 1694.0 (p = 0.765)
192
D.P. Dirette / A comparison of self-generated versus taught internal strategies for working memory Table 3 Mean comparisons of post-test assessments Assessment Face/Name recall CMT GORT-4 CVLT II 1st trial CVLT II 2nd trial RCFT ∗ Significant
Intervention group
Control group
Comparison/value
1.60 (sd = 0.668) 16.07 (sd = 2.51) 17.47 (sd = 6.31) 7.57 (sd = 2.14) 11.86 (sd = 2.17) 37.41 (sd = 8.12)
1.66 (sd = 0.557) 15.30 (sd = 2.62) 18.26 (sd = 6.41) 7.60 (sd = 1.81) 12.12 (sd = 1.89) 34.80 (sd = 6.83)
U = 1718.5 (p = 0.835) U = 1423.5 (p = 0.080) U = 1650.5 (p = 0.596) U = 1696.0 (p = 0.771) U = 1570.5 (p = 0.334) U = 1416.5 (p = 0.075)
difference. Table 4 Mean comparison of follow-up assessments
Assessment Face/Name recall∗ CMT∗ GORT-4 CVLT II 1st trial CVLT II 2nd trial RCFT ∗ Significant
Intervention group
Control group
Comparison/value
2.44 (sd = 1.25) 13.58 (sd = 2.19) 12.58 (sd = 4.40) 9.88 (sd = 2.48) 13.32 (sd = 2.21) 38.81 (7.99)
1.37 (sd = 1.25) 14.61 (sd = 2.52) 14.49 (sd = 6.19) 9.65 (sd = 3.15) 13.78 (sd = 1.86) 38.82 (sd = 5.56)
U = 907.00 (p = 0.000) U = 1255.0 (p = 0.016) U = 1359.5 (p = 0.070) U = 1646.0 (p = 0.807) U = 1516.5 (p = 0.336) U = 1586.5 (p = 0.505)
difference.
vention group participants and did so at the same rate. In addition, the intervention group participants used strategies that were not taught to them at the same rate as the self-generation group participants. Three percent used a motor strategy, 59% used repetition and 10% used strategies that were classified as “other.” None of those strategies were taught during the intervention session. The overall improvement of the groups was similar to that found in previous studies by Derwinger, Stigsdotter Neely and Backman (2005) and by Wang and Thomas (2000). Both groups improved significantly from pretest to follow-up on recall of word lists and on the visual drawing memory test. The control group did not improve significantly on any other tests. The intervention group, however, performed significantly worse on the CMT and the GORT-4 from pre-test to follow-up. Both of these tests were contextual tests. These measures were different from the other tests that included lists, numbers and names. Teaching strategies may be useful for memorization of rote information such as word list, but information that has context, such as stories, may be better memorized using the context that already exists in that information. There were no significant differences in group performance at post-test and only two differences at the one month follow-up. The control group did better on the CMT and the intervention group did better on the Face/Name Recall. The intervention group was taught a specific association strategy for Face/Name Recall that may have helped them improve their performance. The control group may have done better on the CMT
because they were able to use the context of the information or generate their own strategies that were effective for this test. Of the strategies that were taught to the intervention group, none were specifically applicable to the CMT which may have forced the intervention group to try to apply strategies that were not effective. 4.1. Limitations There are a few limitations related to the participants in this study. High functioning students without neurological impairments were both the control and intervention participants, so the study results may be limited in the applicability to populations for whom cognitive rehabilitation is used. Students who participated in the remedial intervention also may have had higher rates of WM internal strategy self-generation than groups who typically receive cognitive rehabilitation. In addition, the control group was significantly older than the intervention group. Although there was equal gender distribution, there were also a limited number of males in the population of both groups. One procedural limitation is the use of group administration for the interventions and assessments. These interventions and assessments may be better used in one-on-one therapy sessions. In addition, as with most studies that use performance measures, normal distributions of the data were difficult to achieve. Therefore, less powerful non-parametric analyses were used. These analyses may have limited ability to find differences as a result of Type II errors.
D.P. Dirette / A comparison of self-generated versus taught internal strategies for working memory
5. Conclusions Participants who generated their own strategies or used a specific strategy for a specific task did as well as participants who were taught myriad internal WM strategies. The participants who were taught strategies demonstrated significant decrease in their performance of contextual WM tasks. Teaching too many strategies may overwhelm participants or may distract them from using the context that is available in such tasks. Teaching strategies in a less repetitious, streamlined format by allowing people to discover strategies on their own and practice the application of those strategies may be more effective for improving contextual memory (Goverover, Arango-Lasprilla, Hillary, Chiaravalloti & DeLuca, 2009). Further research is needed to explore the pace at which internal WM strategies should be taught and to assess the generation and use of strategies by people with neurological impairments. Acknowledgments Thank you to Dr. Michele Anderson for her collaborative contribution to the design and data collection for this study. Also, thank you to Allison Fox, Alex Walcott, Jessica Rector and Kimberlee Hojna for assistance with data collection, test scoring, data entry and literature search support. Conflict of interest This study was funded in part by a Support for Faculty Scholars Award from Western Michigan University. References Anderson, J. (1985). Cognitive Psychology and Its Implications (2nd ed.). New York: W.H. Freeman. Anderson, V. A., Catroppa, C., Morse, S. A., & Haritou, F. (1999). Functional memory skills following traumatic brain injury in young children. Pediatric Rehabilition, 3(4), 159-166. Barker-Collo, S., & Feigin, V. L. (2008). Memory deficit after traumatic brain injury: How big is the problem in New Zealand and what management strategies are available? The New Zealand Medical Journal, 121(1268), 1-11. Chen, C. C., Cermak, S. A., Murray, E. A., & Herderson, A. (1999). The effect of strategy on the recall of the Rey-Osterrieth complex figure in children with or without learning disabilities. Occupational Therapy Journal of Research, 19(4), 258-279. Cicerone, K. D., Dahlberg, C., Kalmar, D., Langenbahn, D. M., Malac, J. F., Bergquist, T. F., Felicetti, T., Giacino, J. T., Harley, J.
193
P., Harrington, D. E., Herzog, J., Kneipp, S., Laatsch, L., & Morse, P. A. (2000). Evidence-based cognitive rehabilitation: Recommendations for clinical practice. Archives of Physical Medicine and Rehabilitation, 81, 1596-1615. Cicerone, K. D., Dahlberg, C., Malec, J. F., Langenbahn, D. M., Felicetti, T., Kneipp, S., Ellmo, W., Kalmar, D., Giacino, J. T., Harley, J. P., Laatsch, L., Morse, P. A., & Catanese, J. (2005). Evidence-based cognitive rehabilitation: Updated review of the literature from 1998 through 2002. Archives of Physical Medicine and Rehabilitation, 86, 1681-1692. Delis, D. C., Kramer, J. H., Kaplan, E., & Ober, B. A. (2000). California Verbal Learning Test (2nd Edition). San Antonio: Psychological Corporation. Derwinger, A., Stigsdotter Neely, A., & Backman, L. (2005). Design your own strategies! Self-generated strategy training versus mnemonic training on old age: An 8-month follow-up. Neuropsychological Rehabilitation: An International Journal, 15(1), 37-54. Doi: 10.1080/09602010343000336 De Wall, C., Wilson, B. A., & Baddeley, A. D. (1994). The Extended Rivermead Behavioural Memory Test: A measure of everyday memory performance in normal adults. Memory, 2(2), 149-166. Diller, L. (1993). An introduction of cognitive rehabilitation. In C. B. Royeen (Ed.), AOTA Self-study Series: Cognitive Rehabilitation, (pp. 7-19). Rockville, MD: American Occupational Therapy Association. Dirette, D. (2012). Progressive neurological disorders. In B. J. Atchison & D. K. Dirette (Eds.), Conditions in Occupational Therapy: Effect on Occupational Performance, (pp. 209-223). Baltimore, MD: Lippincott, Williams and Wilkins. Dombovy, M. L., & Olek, A. C. (1996). Recovery and rehabilitation following traumatic brain injury. Brain Injury, 11(5), 305-318. Goverover, Y., Arango-Lasprilla, J., Hillary, F. G., Chiaravalloti, N., & DeLuca, J. (2009). Application of the spacing effect to improve learning and memory for functional tasks in traumatic brain injury: A pilot study. American Journal of Occupational Therapy, 63(5), 543-548. Kucukdeveci, A. A., Kutlay, S., Elhan, A. H., & Tennant, A. (2008). Construct validity and reliability of the Rivermead Behavioural Memory Test in the Turkish population. Brain Injury, 22(1), 75-82. Lippert-Gruner, M., Kuchta, J., Hellmich, M., & Klug, N. (2006). Neurobehavioral deficits after severe traumatic brain injury (TBI). Brain Injury, 20(6), 569-574. Loring, D. W., Martin, R. C., Meador, K. J., & Lee, G. P. (1990). Psychometric construction of the Rey-Osterrieth complex figure: Methodological considerations and interrater reliability [Electronic version]. Archives of Clinical Neuropsychology, 5(1), 1-14. doi:10.1016/0887-6177(90)90002-7 Luzzi, S., Pesallaccia, M., Fabi, K., Muti, M., Viticchi, G., Provinciali, L., & Piccirilli, M. (2011). Non-verbal memory measured by Rey-Osterrieth complex figure B: Normative data. Neurological Sciences, 32(6), 1081-1089. doi: http://dx.doi.org/10.1007/ s10072- 11-0641-1 McAllister, T. W., Flashman, L. A., McDonald, B. C., & Saykin, A. J. (2006). Mechanisms of working memory dysfunction after mild and moderate TBI: Evidence from functional MRI and neurogenetics. Journal of Neurotrauma, 23(10), 1450-1467. Naismith, S. L., Mowszowski, L., Diamond, K., & Lewis, S. J. G. (2013). Improving memory in Parkinson’s Disease: A healthy brain ageing cognitive training system. Movement Disorders, 28(8), 1097-1103.
194
D.P. Dirette / A comparison of self-generated versus taught internal strategies for working memory
Rasmussen, C., Treit, S., & Pei, J. (2010). Memory interventions for children with memory deficits. In: J. H. Stone & M. Blouin (Eds.), International Encyclopedia of Rehabilitation. Available online: http://cirrie.buffalo.edu/encyclopedia/en/article/276/ Schefft, B. K., Dulay, M. F., & Fargo, J. D. (2008). The use of a selfgeneration memory encoding strategy to improve verbal memory and learning in patients with traumatic brain injury. Applied Neuropsychology, 15, 61-68. DOI: 10.1080/09084280801917806 Smith, S. R., Chang, J., Schnoebelen, K. J., Edwards, J. W., Servesko, A. M., & Walker, S. J. (2007). Psychometrics of a simple method for scoring organizational approach on the Rey-Osterrieth complex figure. Journal of Neuropsychology, 1(1), 39-51. Retrieved January 26, 2014 Stuifbergen, A. K., Becker, H., Perez, F., Morrison, J., Kullberg, V., & Todd, A. (2012). A randomized controlled trial of a cognitive rehabilitation intervention for persons with multiple sclerosis. Clinical Rehabilitation, 26, 882-893. Toglia, J. A. (1993). Contextual Memory Test. San Antonio: Pearson. Toglia, J. A. (1993). Attention and memory. In C.B. Royeen (Ed.), AOTA Self-study Series: Cognitive Rehabilitation, (pp. 1-69). Rockville, MD: American Occupational Therapy Association. Van’t Hooft, I., Andersson, K., Bergman, B., Sejersen, T., von Wendt, L., & Bartfai, A. (2005). Beneficial effect from a cognitive training
program on children with acquired brain injuries demonstrated in a controlled study. Brain Injury, 19(7), 511-518. Wang, A. Y., & Thomas M. H. (2000). Looking for long-term mnemonic effects on serial recall: The legacy of Simonides. American Journal of Psychology, 113(3), 331-340. Wiederholt, J. L., & Bryant, B. R. (2001). Gray Oral Reading Test4th Edition. Austin, TX: Pro-Ed. Woods, S. P., Delis, D. C., Kramer, J. H., & Holdnack, J. H. (2006). The California Verbal Learning Test- second edition: Test-retest, practice effects and reliable change indices for the standard and alternate forms. Archives of Clinical Neuropsychology, 21(5), 413-420. Yassuda, M. S., Flaks, M. K., Viola, L. F., Pereira, F. S., Memoria, C. M., Nunes, P. V., & Forlenza, O. V. (2010). Psychometric characteristics of the Rivermead Behavioural Memory Test (RBMT) as an early detection instrument for dementia and mild cognitive impairment in Brazil. International Psychogeriatrics, 22(6), 1003-1011. Yivisaker, M., Hanks, R., & Johnson-Greene, D. (2002). Perspectives on rehabilitation of individuals with cognitive impairment after brain injury: Rationale for reconsideration of theoretical paradigms. Journal of Head Trauma Rehabilitation, 17(3), 191-209.
NeuroRehabilitation 36 (2015) 187–194 DOI:10.3233/NRE-151206 IOS Press
A comparison of self-generated versus taught internal strategies for working memory Diane Powers Dirette∗ Western Michigan University, 1903 W. Michigan Ave., Kalamazoo, MI 49008, USA Tel.: +1 269 387 7243; Fax: +1 269 387 7262; E-mail: [email protected].
Abstract. BACKGROUND: Internal strategies are effective for improving working memory. These internal working memory strategies can be taught or self-generated. OBJECTIVE: This study compares working memory performance using taught versus self-generated internal working memory strategies and explores the quantity, type and carry-over of the use of these strategies. METHODS: An experimental cohort design with randomly assigned groups compared the performances among 120 participants on 5 memory tests given prior, immediately following and at 1 month post intervention. RESULTS: There were no significant differences in the number or type of internal memory strategies used by the groups with chunking, repetition and association used most commonly. The group that self-generated strategies performed significantly better on a contextual memory test and the group that was taught strategies performed better on a face/name recall test for which a specific strategy was taught. The group that was taught strategies performed significantly worse on contextual memory tests from pre-test to follow-up. CONCLUSIONS: Participants who generated their own strategies or used a specific strategy for a specific task did as well as participants who were taught myriad internal WM strategies. Teaching too many strategies may overwhelm participants or may distract them from using the context that is available in such tasks. Keywords: Cognitive rehabilitation, mnemonics, short-term memory
1. Introduction Working memory (WM) is “the ability to hold information in mind and to manipulate it in light of incoming material” (p. 1450, McAllister, et al., 2006). Because the neural substrates of WM are in regions commonly damaged in traumatic brain injuries (TBI), deficits in WM are common following even mild TBIs (BarkerCollo & Feigin, 2008; McAllister, et al., 2006). Deficits in WM are also common with some progressive neurological disorders such as Mutiple Sclerosis (Dirette, 2012) and Parkinson’s Disease (Naismith, Mowszowski, Diamond & Lewis, 2013). In addition to being a common deficit, WM is resistant to improvement or healing over time, even with intensive restorative therapy (Lippert-Gruner, Kuchta, Hellmich & Klug, 2006). Deficits in WM significantly impact activities
of daily living such as employment, housework, and meal preparation (Dombovy & Olek, 1996), and nearly every aspect of learning (Anderson, Catroppa, Morse & Haritou, 1999). Cognitive rehabilitation is a collaborative multidisciplinary service that often focuses on the improvement of WM. The traditional cognitive rehabilitation paradigm uses restoration and compensatory strategy training methods to improve WM (Ylvisaker, Hanks & Johnson-Greene, 2002). Restorative interventions are typically used first and, if unsuccessful, followed by training in the use of compensatory strategies (Yivisaker, et al., 2002). Compensatory strategies can be either internal or external. Internal WM strategies are techniques used within one’s own mind to enhance the processing and recall of information.
1053-8135/15/$35.00 © 2015 – IOS Press and the authors. All rights reserved
188
D.P. Dirette / A comparison of self-generated versus taught internal strategies for working memory
Some examples of internal memory strategies include chunking, pacing, verbalization, association, acrostics, method of loci, pegword, rehearsal and imagery. Chunking is mentally grouping information into segments to condense the information (Diller, 1993). Pacing is performing a task with intermittent pauses to decrease the effects of proactive interference (Anderson, 1985; Toglia, 1993). Verbalization is orally repeating information gathered from visual sources (Toglia, 1993). Association is creating relationships between the information that is presented and previously learned information. Acrostics are creating expressions using the first letter of each word that needs to be remembered. Method of loci is using association with a known physical layout such as one’s home or office. Pegword is creating a list of words associated with sequential numbers and then using association with the words that were created (e.g. one-fun-shoe with the association that you may love to kick off your shoes and have fun). Rehearsal is mentally repeating the information that was presented. And, imagery is creating a mental picture of information that is presented orally. The term “mnemonics” is often used in the literature to describe internal WM strategies. Mnemonics are any techniques used to help remember information. Short rhymes or phrases that elaborate on the information are often thought of when the term mnemonics is used. Repetition or rehearsal of information is not generally considered a mnemonic as there is an implication in mnemonics that something about the information is elaborated on to help with encoding and recall. 1.1. Literature review Although cognitive rehabilitation has been shown to be effective for improving WM through the use of compensatory strategies (Cicerone, Dahlberg, Kalmer, et al., 2000; Cicerone, Dahlberg, Malec et al., 2005), the most effective means for developing the use of these strategies has not been established (Rasmussen, Treit, & Pei, 2010). Internal memory strategies can be taught by therapists or people may develop their own strategies. The development of strategies without instruction is referred to as self-generation. Cognitive rehabilitation therapists typically teach people myriad strategies and allow them to practice and choose the ones that work the best for them. According to the generation effect, however, optimal acquisition and retention of information is best established though active participation rather than by passive observation (Schefft, Dulay, & Fargo, 2008). It is unknown if this effect applies to
self-generation of internal strategies or if the traditional method of teaching strategies is more effective. Studies typically compare an intervention that teaches compensatory strategies to an intervention that is deemed restorative or remedial. In the remedial intervention, the participants are asked to repeatedly perform a cognitive task, but they are not given instruction in the use of strategies to complete the task. For example, a randomized, controlled trial of 61 adults who had Multiple Sclerosis with mild cognitive deficits, compared an intervention focused on training in the use of strategies to a remedial intervention (Stuifbergen, Becker, Perez, Morrison, Kullberg, & Todd, 2012). Participants were tested initially and at the end of treatment using The Minimal Assessment of Cognitive Function. Both groups improved significantly and there were no significant differences in performance except that the strategy group did significantly better on a test of verbal memory. The use of self-generated strategies by the remedial group, however, may have confounded the results as there was no measure of strategy use by the remedial group. A couple of studies have compared WM performance for taught versus self-generated strategy groups. Derwinger, Stigsdotter Neely and Backman (2005) compared the memory performance of 28 older adults (61–83 years old) who were taught a mnemonic for recall of a 4 digit number to that of 29 age-matched participants who were encouraged to use self-generated strategies to recall the same numbers. The study also included a control group of 24 participants who received no training. All three groups improved equally on memory tests during the study. In the long-term follow-up testing, however, the performance of the mnemonic group decreased, the performance of the self-generated group improved, and the performance of the control group stayed the same. The researchers concluded that self-generated strategies may be more beneficial to older adults than a taught strategy because selfgenerated strategies do not require practice to maintain their use. Wang and Thomas (2000) explored the differences in serial word recall among 202 adult college students who were divided into three groups, two mnemonic groups and one control group. The two mnemonic groups included one group who was taught the method of loci and one group who was taught the pegword system. The control group was told to use their own best strategies. After a two-day delay, the three groups demonstrated no significant differences in their ability to recall the serial word lists.
D.P. Dirette / A comparison of self-generated versus taught internal strategies for working memory
These studies provide preliminary exploration of the differences in self-generated versus taught internal strategies for WM, but they offer limited exploration of the carry-over of the use of the strategies, the types of strategies that were used by the groups or the types of tasks that may be enhanced by the strategies that are used. The participants who were taught strategies may have added self-generated strategies to the process, as well. The purpose of this current study is to compare the immediate and long-term performance that results from taught versus self-generated internal WM strategies and to compare the quantity and type of strategies that are used by the two groups.
2. Method 2.1. Design An experimental cohort design was used to examine the differences in internal WM strategy use with groups randomly assigned to be either control or intervention groups. Data were not collected simultaneously. Participants were blind to their group assignment and the test scoring was completed by a research assistant who was blind to group participation. 2.2. Participants All participants were students at the university at which the study was conducted. HSIRB approval was obtained for the recruitment and study procedures. Participants were recruited at the beginning of a classroom session without the instructor present and given the option to participate in the study or leave the classroom. The instructors were never given any information about attendance or participation. All participants provided individual informed consent and completed a demographic data sheet. The testing and intervention sessions were completed in a group format. 2.3. Procedures Pre-tests were given at the beginning of each session for all groups before any interventions were provided. Immediately following the pre-tests, the intervention group participated in a 30 minute training session on the use of internal WM strategies. The control group participated in the same length of intervention time by practicing the same activities that were used with the intervention group. The control group, however,
189
received no training in the use of WM strategies during the activities. The internal WM strategies that were taught to the intervention group during this study included chunking, pacing, verbalization, association, acrostics, method of loci, peg word, rehearsal and imagery. These 9 strategies were taught in the order listed. Each strategy was taught using an activity. Chunking was taught using a series of numbers such as those for a phone number or credit card, pacing was taught using a short story, verbalization was taught using addresses and phone numbers, association was taught using names of people and places and acrostics was taught using a series of words such as the Great Lakes, the cranial nerves and the planets. Post-tests were given immediately after the intervention and repeated at follow-up, one month following the intervention. Different versions of each test were given at each testing session to reduce the risk of a practice effect. A standardized order and script were used during all testing sessions. The participants were not allowed to write down any information, use electronic devices or make any notations during any of the assessments. In other words, the use of external WM strategies was not allowed. At the end of the follow-up session, the participants completed a brief questionnaire about their use of internal memory strategies during the testing procedures. They were asked to identify the number and types of strategies that were used. 2.4. Instruments Five memory tests were given to each group at pre-test, post-test and follow-up. The five tests included Face/Name Recall, the Contextual Memory Test (CMT), the Grey Oral Reading Test, 4th Edition (GORT-4), the California Verbal Learning Test (CVLT II) and the Rey-Osterrieth Complex Figure Test (ROCF). Face/Name Recall was given using two faces/names at pre-test, and four faces/names at post-test and followup. The pictures, names and procedures that were used were similar to those used in the Rivermead Behavioural Memory Test (RBMT) with a black and white picture of a face shown for five seconds on a document camera and the first and last name spoken at the beginning of that visual presentation. The participants were then asked to recall the names of the people when shown the faces after a 20 minute delay. The RBMT has been shown to have good reliability and validity
190
D.P. Dirette / A comparison of self-generated versus taught internal strategies for working memory
with varied populations (DeWall, Wilson, & Braddeley, 1994; Kucukdeveci, Kutlay, Elhan, & Tennant, 2008; Yassuda, et al., 2010). The CMT is a standardized test of visual memory that uses a picture card of 20 related, everyday objects. The picture card is presented for 90 seconds and then the participant is asked to recall as many objects as possible. The participants in this study were asked to write these objects on a list instead of recalling them verbally. Test-Retest reliability scores for the CMT range from 0.74 to 0.94 (Toglia, 1993). The concurrent validity with the RBMT screening scores range from 0.80 to 0.85 (Toglia, 1993). Paragraphs 8, 9 and 10 from the GORT-4 (Wiederholt & Bryant, 2001) were used for pre-tests, post-tests and follow-up tests respectively to assess recall of written information. The participants were given time to read the paragraphs to the end and then asked to recall as much information as possible. Recall was scored using the measurement of information bits written down by the participant immediately following reading the paragraph. Order of the story recall was not considered. The reliability and validity measures for the GORT-4 do not apply to this study as the test was not used with the test questions, but used as a measure of information recall as it is done in the RBMT. The CVLT II was administered as a measure of auditory recall. The list of 16 words was read to the participants at an interval of one word per second. The participants were asked to immediately recall and write down as many items from the list as they could remember. As told to the participants, order was not considered. The same list was then repeated in the same order and at the same rate. After the list was read for the second time, the participants were asked to write down on a new piece of paper as many items as they could remember the second time. The test/re-test reliability scores for the CVLT II range from 0.80–0.84 for the same version of the test and 0.61–0.73 for an alternative version of the test (Woods, Delis, Dramer, & Holdnack, 2006). Internal consistency for the CVLT II ranges from.78 to.94 (Delis, Kramer, Kaplan, & Ober, 2000). The ROCF is an assessment of visuospatial recall in which participants are asked to reproduce a complicated line drawing. The ROCF was scored using the formal scoring criteria for the standard 18 elements with 0.5–2.0 points awarded for each reproduced element depending on accuracy/distortion and location of the reproduction. The ROCF has been shown to have good validity (Smith, et al., 2007) and inter-rater reli-
ability scores ranging from 0.92–0.97 (Loring, Martin, Meador, & Lee, 1990; Luzzi, et al., 2001). Figure A was used for pre-tests and follow-up tests and Figure B was used for post-tests. 2.5. Data analyses All data were entered into SPSS version 17 for the analysis. Variables were analyzed for normal distributions. Where skewed variables occurred, a non-parametric analysis was used to make group comparisons. All other group comparisons were completed using t-tests or RMANOVA as appropriate. In cases where tests of homogeneity were not met, the equality not assumed or the Greenhouse-Giesser correction was used. Alpha was set at 0.05 as is appropriate for comparing groups of this size.
3. Results 3.1. Demographics Altogether there were 120 participants in the study with a mean age of 22.24 years (sd = 5.61) with a range of 18–58 years. There were 50 participants in the control group and 70 participants in the intervention group. The mean age of the control group was 23.81 years (sd = 6.52) with a range of 19–58 years. The mean age of the intervention group was 21.19 years (sd = 4.66) with a range of 18–50 years. The age of the control group was significantly higher (U = 967.5, sign. = 0.000). All of the participants were female except for four male participants in the control group and five male participants in the intervention group. The assessments and interventions were all completed in three control and three intervention groups. The size of the groups ranged from 11 to 28 participants. The sizes of the control groups were 11, 16 and 23 participants and the sizes of the intervention groups were 18, 24 and 28 participants. 3.2. Strategy use There was no significant difference (t = 0.886; sign = 0.378) between the total number of internal WM strategies used by the control group (X = 2.76; sd = 0.990) and by the intervention group (X = 2.59; sd = 0.960). A series of chi square analyses also revealed no significant differences in the types of internal WM strategies that were used by the groups. Chunking was the strategy used most often by both groups with 83%
D.P. Dirette / A comparison of self-generated versus taught internal strategies for working memory
191
of the intervention group and 80% of the control group reportedly employing this strategy. The second most commonly used strategy was repetition with 59% of the intervention group and 64% of the control group using this strategy. The third most common was association (39% of the intervention group and 46% of the control group), the forth was visualization (19% of the intervention group and 30% of the control group), the fifth was acrostics (17% of the intervention group and 20% of the control group) and the sixth was verbalization (14% of the intervention group and 4% of the control group). The seventh through tenth most commonly used strategies were method of loci, motor practice, pacing, and pegging with fewer than 10% of participants using those strategies in either group. One category of strategy use was deemed “other”. There was no significant difference in this category with 10% of the intervention group and 20% of the control group reportedly using a strategy that was not captured in the listed categories. The most common of the “other” strategies included sequencing and counting.
performance was significantly worse on these two WM measures. See Table 1. There was one significant difference between group scores at pre-test. The control group did significantly better on the CVLT II trial 1 (U = 1276.5, p = 0.011) and trial 2 (U = 1278.5, p = 0.011). The mean scores for the intervention group were trial 1 x = 7.09 (sd = 1.99) and trial 2 x = 11.33 (sd = 2.09). The mean scores for the control group were trial 1 x = 7.92 (sd = 1.82) and trial 2 x = 12.22 (sd = 2.21). See Table 2. There were no significant differences between group scores at post-test. See Table 3. There were only two significant differences between the groups at the one month follow-up. The control group did significantly better on the follow-up test of the CMT (U = 1255.0, p = 0.016) and the intervention group did significantly better on the follow-up text of Face/Name Recall (U = 907.0, p = 0.000). See Table 4.
3.3. Comparisons of scores
There were no differences in the number or type of internal WM strategies used by the participants who were taught (intervention group) and the participants who were given a remedial intervention during which they self-generated internal WM strategies (control group). Participants from both groups used between 2 and 3 different types of strategies during the follow-up testing. The control group participants self-generated and used the same strategies that were used by the inter-
Both groups improved significantly on the CVLT (trials 1 and 2) and the RCFT from pre-test to follow-up testing. The intervention group also improved significantly on Face/Name Recall, but the control group’s performance did not change significantly. The control group’s performance did not change significantly on the CMT or the GORT-4, but the intervention group’s
4. Discussion
Table 1 Comparisons of pre-test to follow-up scores Assessment
Intervention group t = 6.023 (p = 0.000)∗ t = 3.644 (p = 0.001)∗∗ t = 3.024 (p = 0.004)∗∗ t = 9.120 (p = 0.000)∗ t = 7.430 (p = 0.000)∗ t = 3.643 (p = 0.001)∗
Face/Name recall CMT GORT-4 CVLT II 1st trial CVLT II 2nd trial RCFT ∗ Significantly
Control group t = 1.512 (p = 0.137) t = 0.339 (p = 0.691) t = 2.001 (p = 0.051) t = 4.107 (p = 0.000)∗ t = 3.848 (p = 0.000)∗ t = 3.919 (p = 0.000)∗
better. ∗∗ Significantly worse. Table 2 Mean comparisons of pre-test assessments
Assessment Face/Name recall CMT GORT-4 CVLT II 1st trial∗ CVLT II 2nd trial∗ RCFT ∗ Significant
difference.
Intervention group
Control group
Comparison/value
1.55 (sd = 0.654) 14.69 (sd = 2.56) 14.37 (sd = 5.27) 7.09 (sd = 1.99) 11.33 (sd = 2.09) 35.87 (sd = 7.72)
1.66 (sd = 0.626) 14.72 (sd = 2.17) 16.06 (sd = 4.91) 7.92 (sd = 1.82) 12.22 (sd = 2.21) 35.36 (sd = 6.74)
U = 1559.5 (p = 0.277) U = 1749.0 (p = 0.996) U = 1441.0 (p = 0.099) U = 1276.5 (p = 0.011) U = 1278.5 (p = 0.011) U = 1694.0 (p = 0.765)
192
D.P. Dirette / A comparison of self-generated versus taught internal strategies for working memory Table 3 Mean comparisons of post-test assessments Assessment Face/Name recall CMT GORT-4 CVLT II 1st trial CVLT II 2nd trial RCFT ∗ Significant
Intervention group
Control group
Comparison/value
1.60 (sd = 0.668) 16.07 (sd = 2.51) 17.47 (sd = 6.31) 7.57 (sd = 2.14) 11.86 (sd = 2.17) 37.41 (sd = 8.12)
1.66 (sd = 0.557) 15.30 (sd = 2.62) 18.26 (sd = 6.41) 7.60 (sd = 1.81) 12.12 (sd = 1.89) 34.80 (sd = 6.83)
U = 1718.5 (p = 0.835) U = 1423.5 (p = 0.080) U = 1650.5 (p = 0.596) U = 1696.0 (p = 0.771) U = 1570.5 (p = 0.334) U = 1416.5 (p = 0.075)
difference. Table 4 Mean comparison of follow-up assessments
Assessment Face/Name recall∗ CMT∗ GORT-4 CVLT II 1st trial CVLT II 2nd trial RCFT ∗ Significant
Intervention group
Control group
Comparison/value
2.44 (sd = 1.25) 13.58 (sd = 2.19) 12.58 (sd = 4.40) 9.88 (sd = 2.48) 13.32 (sd = 2.21) 38.81 (7.99)
1.37 (sd = 1.25) 14.61 (sd = 2.52) 14.49 (sd = 6.19) 9.65 (sd = 3.15) 13.78 (sd = 1.86) 38.82 (sd = 5.56)
U = 907.00 (p = 0.000) U = 1255.0 (p = 0.016) U = 1359.5 (p = 0.070) U = 1646.0 (p = 0.807) U = 1516.5 (p = 0.336) U = 1586.5 (p = 0.505)
difference.
vention group participants and did so at the same rate. In addition, the intervention group participants used strategies that were not taught to them at the same rate as the self-generation group participants. Three percent used a motor strategy, 59% used repetition and 10% used strategies that were classified as “other.” None of those strategies were taught during the intervention session. The overall improvement of the groups was similar to that found in previous studies by Derwinger, Stigsdotter Neely and Backman (2005) and by Wang and Thomas (2000). Both groups improved significantly from pretest to follow-up on recall of word lists and on the visual drawing memory test. The control group did not improve significantly on any other tests. The intervention group, however, performed significantly worse on the CMT and the GORT-4 from pre-test to follow-up. Both of these tests were contextual tests. These measures were different from the other tests that included lists, numbers and names. Teaching strategies may be useful for memorization of rote information such as word list, but information that has context, such as stories, may be better memorized using the context that already exists in that information. There were no significant differences in group performance at post-test and only two differences at the one month follow-up. The control group did better on the CMT and the intervention group did better on the Face/Name Recall. The intervention group was taught a specific association strategy for Face/Name Recall that may have helped them improve their performance. The control group may have done better on the CMT
because they were able to use the context of the information or generate their own strategies that were effective for this test. Of the strategies that were taught to the intervention group, none were specifically applicable to the CMT which may have forced the intervention group to try to apply strategies that were not effective. 4.1. Limitations There are a few limitations related to the participants in this study. High functioning students without neurological impairments were both the control and intervention participants, so the study results may be limited in the applicability to populations for whom cognitive rehabilitation is used. Students who participated in the remedial intervention also may have had higher rates of WM internal strategy self-generation than groups who typically receive cognitive rehabilitation. In addition, the control group was significantly older than the intervention group. Although there was equal gender distribution, there were also a limited number of males in the population of both groups. One procedural limitation is the use of group administration for the interventions and assessments. These interventions and assessments may be better used in one-on-one therapy sessions. In addition, as with most studies that use performance measures, normal distributions of the data were difficult to achieve. Therefore, less powerful non-parametric analyses were used. These analyses may have limited ability to find differences as a result of Type II errors.
D.P. Dirette / A comparison of self-generated versus taught internal strategies for working memory
5. Conclusions Participants who generated their own strategies or used a specific strategy for a specific task did as well as participants who were taught myriad internal WM strategies. The participants who were taught strategies demonstrated significant decrease in their performance of contextual WM tasks. Teaching too many strategies may overwhelm participants or may distract them from using the context that is available in such tasks. Teaching strategies in a less repetitious, streamlined format by allowing people to discover strategies on their own and practice the application of those strategies may be more effective for improving contextual memory (Goverover, Arango-Lasprilla, Hillary, Chiaravalloti & DeLuca, 2009). Further research is needed to explore the pace at which internal WM strategies should be taught and to assess the generation and use of strategies by people with neurological impairments. Acknowledgments Thank you to Dr. Michele Anderson for her collaborative contribution to the design and data collection for this study. Also, thank you to Allison Fox, Alex Walcott, Jessica Rector and Kimberlee Hojna for assistance with data collection, test scoring, data entry and literature search support. Conflict of interest This study was funded in part by a Support for Faculty Scholars Award from Western Michigan University. References Anderson, J. (1985). Cognitive Psychology and Its Implications (2nd ed.). New York: W.H. Freeman. Anderson, V. A., Catroppa, C., Morse, S. A., & Haritou, F. (1999). Functional memory skills following traumatic brain injury in young children. Pediatric Rehabilition, 3(4), 159-166. Barker-Collo, S., & Feigin, V. L. (2008). Memory deficit after traumatic brain injury: How big is the problem in New Zealand and what management strategies are available? The New Zealand Medical Journal, 121(1268), 1-11. Chen, C. C., Cermak, S. A., Murray, E. A., & Herderson, A. (1999). The effect of strategy on the recall of the Rey-Osterrieth complex figure in children with or without learning disabilities. Occupational Therapy Journal of Research, 19(4), 258-279. Cicerone, K. D., Dahlberg, C., Kalmar, D., Langenbahn, D. M., Malac, J. F., Bergquist, T. F., Felicetti, T., Giacino, J. T., Harley, J.
193
P., Harrington, D. E., Herzog, J., Kneipp, S., Laatsch, L., & Morse, P. A. (2000). Evidence-based cognitive rehabilitation: Recommendations for clinical practice. Archives of Physical Medicine and Rehabilitation, 81, 1596-1615. Cicerone, K. D., Dahlberg, C., Malec, J. F., Langenbahn, D. M., Felicetti, T., Kneipp, S., Ellmo, W., Kalmar, D., Giacino, J. T., Harley, J. P., Laatsch, L., Morse, P. A., & Catanese, J. (2005). Evidence-based cognitive rehabilitation: Updated review of the literature from 1998 through 2002. Archives of Physical Medicine and Rehabilitation, 86, 1681-1692. Delis, D. C., Kramer, J. H., Kaplan, E., & Ober, B. A. (2000). California Verbal Learning Test (2nd Edition). San Antonio: Psychological Corporation. Derwinger, A., Stigsdotter Neely, A., & Backman, L. (2005). Design your own strategies! Self-generated strategy training versus mnemonic training on old age: An 8-month follow-up. Neuropsychological Rehabilitation: An International Journal, 15(1), 37-54. Doi: 10.1080/09602010343000336 De Wall, C., Wilson, B. A., & Baddeley, A. D. (1994). The Extended Rivermead Behavioural Memory Test: A measure of everyday memory performance in normal adults. Memory, 2(2), 149-166. Diller, L. (1993). An introduction of cognitive rehabilitation. In C. B. Royeen (Ed.), AOTA Self-study Series: Cognitive Rehabilitation, (pp. 7-19). Rockville, MD: American Occupational Therapy Association. Dirette, D. (2012). Progressive neurological disorders. In B. J. Atchison & D. K. Dirette (Eds.), Conditions in Occupational Therapy: Effect on Occupational Performance, (pp. 209-223). Baltimore, MD: Lippincott, Williams and Wilkins. Dombovy, M. L., & Olek, A. C. (1996). Recovery and rehabilitation following traumatic brain injury. Brain Injury, 11(5), 305-318. Goverover, Y., Arango-Lasprilla, J., Hillary, F. G., Chiaravalloti, N., & DeLuca, J. (2009). Application of the spacing effect to improve learning and memory for functional tasks in traumatic brain injury: A pilot study. American Journal of Occupational Therapy, 63(5), 543-548. Kucukdeveci, A. A., Kutlay, S., Elhan, A. H., & Tennant, A. (2008). Construct validity and reliability of the Rivermead Behavioural Memory Test in the Turkish population. Brain Injury, 22(1), 75-82. Lippert-Gruner, M., Kuchta, J., Hellmich, M., & Klug, N. (2006). Neurobehavioral deficits after severe traumatic brain injury (TBI). Brain Injury, 20(6), 569-574. Loring, D. W., Martin, R. C., Meador, K. J., & Lee, G. P. (1990). Psychometric construction of the Rey-Osterrieth complex figure: Methodological considerations and interrater reliability [Electronic version]. Archives of Clinical Neuropsychology, 5(1), 1-14. doi:10.1016/0887-6177(90)90002-7 Luzzi, S., Pesallaccia, M., Fabi, K., Muti, M., Viticchi, G., Provinciali, L., & Piccirilli, M. (2011). Non-verbal memory measured by Rey-Osterrieth complex figure B: Normative data. Neurological Sciences, 32(6), 1081-1089. doi: http://dx.doi.org/10.1007/ s10072- 11-0641-1 McAllister, T. W., Flashman, L. A., McDonald, B. C., & Saykin, A. J. (2006). Mechanisms of working memory dysfunction after mild and moderate TBI: Evidence from functional MRI and neurogenetics. Journal of Neurotrauma, 23(10), 1450-1467. Naismith, S. L., Mowszowski, L., Diamond, K., & Lewis, S. J. G. (2013). Improving memory in Parkinson’s Disease: A healthy brain ageing cognitive training system. Movement Disorders, 28(8), 1097-1103.
194
D.P. Dirette / A comparison of self-generated versus taught internal strategies for working memory
Rasmussen, C., Treit, S., & Pei, J. (2010). Memory interventions for children with memory deficits. In: J. H. Stone & M. Blouin (Eds.), International Encyclopedia of Rehabilitation. Available online: http://cirrie.buffalo.edu/encyclopedia/en/article/276/ Schefft, B. K., Dulay, M. F., & Fargo, J. D. (2008). The use of a selfgeneration memory encoding strategy to improve verbal memory and learning in patients with traumatic brain injury. Applied Neuropsychology, 15, 61-68. DOI: 10.1080/09084280801917806 Smith, S. R., Chang, J., Schnoebelen, K. J., Edwards, J. W., Servesko, A. M., & Walker, S. J. (2007). Psychometrics of a simple method for scoring organizational approach on the Rey-Osterrieth complex figure. Journal of Neuropsychology, 1(1), 39-51. Retrieved January 26, 2014 Stuifbergen, A. K., Becker, H., Perez, F., Morrison, J., Kullberg, V., & Todd, A. (2012). A randomized controlled trial of a cognitive rehabilitation intervention for persons with multiple sclerosis. Clinical Rehabilitation, 26, 882-893. Toglia, J. A. (1993). Contextual Memory Test. San Antonio: Pearson. Toglia, J. A. (1993). Attention and memory. In C.B. Royeen (Ed.), AOTA Self-study Series: Cognitive Rehabilitation, (pp. 1-69). Rockville, MD: American Occupational Therapy Association. Van’t Hooft, I., Andersson, K., Bergman, B., Sejersen, T., von Wendt, L., & Bartfai, A. (2005). Beneficial effect from a cognitive training
program on children with acquired brain injuries demonstrated in a controlled study. Brain Injury, 19(7), 511-518. Wang, A. Y., & Thomas M. H. (2000). Looking for long-term mnemonic effects on serial recall: The legacy of Simonides. American Journal of Psychology, 113(3), 331-340. Wiederholt, J. L., & Bryant, B. R. (2001). Gray Oral Reading Test4th Edition. Austin, TX: Pro-Ed. Woods, S. P., Delis, D. C., Kramer, J. H., & Holdnack, J. H. (2006). The California Verbal Learning Test- second edition: Test-retest, practice effects and reliable change indices for the standard and alternate forms. Archives of Clinical Neuropsychology, 21(5), 413-420. Yassuda, M. S., Flaks, M. K., Viola, L. F., Pereira, F. S., Memoria, C. M., Nunes, P. V., & Forlenza, O. V. (2010). Psychometric characteristics of the Rivermead Behavioural Memory Test (RBMT) as an early detection instrument for dementia and mild cognitive impairment in Brazil. International Psychogeriatrics, 22(6), 1003-1011. Yivisaker, M., Hanks, R., & Johnson-Greene, D. (2002). Perspectives on rehabilitation of individuals with cognitive impairment after brain injury: Rationale for reconsideration of theoretical paradigms. Journal of Head Trauma Rehabilitation, 17(3), 191-209.
