A simultaneous examination of two forms of working memory training: Evidence for near transfer only

May combine two claims.

The verbal complex span training produced no evidence of transfer to verbal short-term memory or span tasks despite robust training gains.

“We begin with the verbal complex span training as we were surprised to see no evidence for transfer to either verbal short-term memory or span tasks in spite of the robust training gains”

May combine two claims.

The average maximum n-back level reached at any point was 6.1, with an average gain of 2.5 levels.

“The average maximum level reached during training at any point was 6.1 with an average gain of 2.5 levels (from lowest n-back level to highest).”

Probably not a substantial claim: may not pass S3 (load-bearing).

Among the best n-back performers, no single strategy predominated.

“Amongst our best performers there was no one strategy that predominated.”

Probably not a substantial claim: may not pass S3 (load-bearing).

The only personality measure showing a between-group difference approaching significance was the Adult Hope Scale, F(3, 116) = 2.59, p = .056, eta_p2 = .07.

“The only difference between any of the groups that approached significance was on the Adult Hope Scale, F (3, 116) = 2.59, p = .056,η p 2 = .07.”

May combine two claims.

All transfer tasks were significantly correlated with at least one other measure in their proposed construct, with two exceptions.

“All tasks were significantly correlated with at least one other measure in their proposed construct with two exceptions.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Video game scores ranged from 700 to 123,580 at Session 1 and from 34,896 to 305,840 by Session 12.

“This score ranged from 700 to 123, 580 points at Session 1 and from 34,896 to 305,840 points by Session 12.”

May combine two claims.

There were group differences in reported motivation, F(3, 114) = 2.96, p = .04, eta_p2 = .08, and perceived challenge, F(3, 114) = 16.9, p < .001, eta_p2 = .32, with the non-adaptive n-back group reporting less motivation and challenge.

“We found group differences both in reported motivation, F (3,114) = 2.96, p = .04,η p 2 = .08and how challenging the training was perceived, F (3,114) = 16.9, p <.001, η p 2 = .32with the non-adaptive n-back group reporting being less motivated and challenged.”

May combine two claims.

Graphical examination showed overall pre-to-post improvement on both n-back transfer tasks for all groups except the RTS group on verbal n-back, with spatial n-back adaptive and non-adaptive groups showing larger increases than RTS and verbal span groups.

“A graphical examination revealed the same pattern for both tasks where overall all groups improved their performance from pre-to post-test (with the exception of the RTS group on the verbal n-back) with the spatial n-back adaptive and non-adaptive training groups showing even larger increases than the RTS and verbal span training groups.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Only one verbal complex span participant reported an explicitly visual strategy combining words into a visual image.

“Only one participant reported an explicitly visual strategy in which words were assigned to the letters and then combined into a visual image.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

All verbal complex span participants reported some chunking strategy, with the most common being assigning a word to each letter and creating sentences; 24 explicitly reported this, and the rest reported chunking into initials or acronyms.

“For verbal complex span, all of our participants reported some sort of chunking strategy for remembering the letters with the most popular being the assignment of a word to each letter and creating sentences with those words.Twenty-four of our participants explicitly reported this strategy while the rest of our participants reported either chunking the letters into people's initials or acronyms.”

May combine two claims.

The verbal complex span training gains were equivalent to or greater than previous studies, with a 36% increase compared with the 25% increase reported by Chein and Morrison (2010).

“in which our participants showed improvements equivalent to or greater than previous studies (36 % increase compared to the 25 % increase reported in Chein & Morrison (2010)).”

• cites Chein & Morrison (2010) sourceNot checked yet

  Not checked in the free run.

May combine two claims.

The only clear transfer observed was near transfer from the spatial n-back training to novel n-back tasks, and this occurred for both adaptive and non-adaptive n-back.

“We only found clear evidence for near transfer from the spatial n-back training to new forms of n-back, and this was the case for both adaptive and non-adaptive n-back.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Most participants showed considerable session-to-session fluctuation in n-back level.

“However, we did see a great deal of fluctuation for most participants in n-back level achieved from session to session.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Participants reported being fairly motivated, with an average score of 3.8 out of 5.

“Overall participants reported being fairly motivated with an average score of 3.8 out of 5.”

May combine two claims.

One video game participant showed no or negative improvement, with a high score of 123,580 on the first day and an average of 46,000 across remaining days.

“However, only one individual showed no or negative improvement with a high score of 123,580 on the first day with notably worse performance with an average score of 46, 000 across the remaining days of training.”

May combine two claims.

For the remaining composites, Bayes factors were all greater than 3, indicating support for the null.

“For the remaining composites, the Bayes factors were all >3 indicating support for the null.”

Probably not a substantial claim: may not pass S3 (load-bearing).

All video game participants other than the dropped one showed positive improvement in game score across training.

“All other participants demonstrated positive improvements in game score across training sessions.”

May combine two claims.

For n-back transfer tasks, there was a significant interaction for Verbal n-back, F(3, 112) = 5.6, p < .001, eta_p2 = .13, but not for Object n-back, F(3, 112) = 2.2, p = .10, eta_p2 = .05.

“For the n-back transfer tasks, we found a significant interaction for the Verbal n-back task, F (3, 112) = 5.6, p < .001,η p 2 = .13,but not for the Object n-back task, F (3, 112) = 2.2, p =.10, η p 2 = .05.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Among 31 spatial n-back participants, maximum n-back levels varied: three peaked at 4-back, six at 5-back, 12 at 6-back, seven at 7-back, one at 8-back, and two at 9-back.

“Of our 31 participants, three peaked at a maximum level of 4-back, six at 5-back, 12 at 6-back, seven at 7-back, one at 8-back, and two at 9-back.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Eleven n-back participants reported counting as sole or combined strategy; three assigned a noise or musical note to each location; seven reported no clear strategy.

“Eleven participants reported counting either as the sole strategy or in combination with tapping or drawing shapes.Three participants reported assigning a noise or musical note to each location and trying to keep track of the sequence that way.Finally, seven participants did not report a clear strategy or simply wrote they had no strategy.”

Probably not a substantial claim: may not pass S3 (load-bearing).

A majority of participants in each group endorsed recommending their training to others.

“and a majority of participants in each group endorsed recommending the training they participated in to others.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

In n-back training, strategies were diverse; the most common was using stimulus locations to draw shapes or patterns in the mind's eye, reported by 12 participants alone or with counting, tapping, or tracing.

“The most common strategy was to use the locations where the stimuli appeared to draw shapes or patterns in the mind's eye.Twelve of our participants reported using this method either alone or in combination with counting the number of squares that had appeared as well as tapping or tracing out a corresponding grid on their leg or on the table.”

The quote only partly states this claim.

The non-adaptive n-back group showed similar improvement on letter and object n-back transfer tasks as the adaptive n-back group.

“We also saw a similar improvement on the letter and object n-back transfer tasks for the non-adaptive n-back group”

May combine two claims.

The study found similar relationships between neuroticism and training (r = -.25) and self-reported stress and training (r = -.26).

“We did see a similar relationship between neuroticism and training (r =-.25) as well as between self-reported stress and training (r = -.26).”

May combine two claims.

The few correlations between training and personality variables were small and underpowered.

“similar to previous studies, the few correlations seen between training and personality variables were small and underpowered.”

May combine two claims.

The study found no relationships between training and conscientiousness or related variables such as need for cognition, grit, or self-control, nor between personality and near transfer.

“However, we did not see any relationships between training and conscientiousness or related variables such as need for cognition, grit, or self-control nor evidence of a relationship between personality and the near transfer observed.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Both n-back and complex span tasks predict variance in fluid intelligence scores.

“Both tasks are deemed working-memory tasks because they require the simultaneous processing and storage of information and both predict variance in fluid intelligence scores (Kane, Conway, Miura, & Colflesh, 2007).”

• cites Kane, Conway, Miura, & Colflesh, 2007Not checked yet

  Not checked in the free run.

Classified as probably not a substantial claim.

Participants were randomly assigned to one of four conditions at the start of the study.

“After agreeing to participate in the study, participants were initially randomly assigned to one of the four conditions.”

May combine two claims.

A meta-analysis found only a weak relationship between n-back and complex span tasks and concluded that they tap different components of working memory.

“However, a recent meta-analysis of studies measuring the relationship between the n-back and complex span found only a weak relationship between the two tasks and concluded they tap different components of working memory (Redick & Lindsey, 2013).”

• cites Redick & Lindsey, 2013 sourceNot checked yet

  Not checked in the free run.

May combine two claims. The quote only partly states this claim.

Post hoc power analysis indicated the sample size was large enough to detect Group by Testing time interactions; with average pre-post correlation .55, power to detect medium d = .5 or large d = .8 was over .99, and with lowest correlation r = .14, power was .94.

“Using the average correlation between our pre-and post-test measurements (average r =. 55), our power to detect a medium (Cohen's d =.5) or a large effect (Cohen's d =.8) was over .99 and using our lowest correlation r =.14 (ANT), power was .94.”

Probably not a substantial claim: may not pass S3 (load-bearing).

The planned sample size was based on the medium effect size of Cohen's d = .65 reported by Jaeggi et al. (2008).

“We based our planned sample sizes on the medium effect size (Cohen's d = .65)reported by Jaeggi et al. (2008) using a similar sample size.”

• cites Jaeggi et al. (2008) sourceNot checked yet

  Not checked in the free run.

Classified as probably not a substantial claim.

Multiple measures of each construct were collected so that evidence of transfer would not rely on a single task.

“We collected multiple measures of each construct so that evidence of transfer would not rely upon a single task and when possible tried to choose tasks that had shown prior evidence of being sensitive to some form of training intervention.”

Classified as probably not a substantial claim.

Single n-back training, using only spatial or only verbal stimuli, has been reported to be as effective as dual n-back on measures of spatial reasoning.

“In the case of the n-back, single n-back training in which the only spatial or only verbal stimuli are used has been reported as being as effective as the dual n-back (combining verbal and spatial) on measures of spatial reasoning (Jaeggi, Studer-Luethi, Buschkuehl, Su, Jonides, & Perrig, 2010).”

• cites Jaeggi, Studer-Luethi, Buschkuehl, Su, Jonides, & Perrig, 2010 sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing).

An active control group is preferred over a no-contact control group in cognitive training studies.

“While some studies have employed a simple no-contact control group in which participants are simply tested on the transfer measures twice with no other interaction with experimenters, an active control group in which participants undergo some form of intervention is preferred.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

The final sample had 34, 35, 33, and 30 participants assigned to spatial n-back, verbal complex span, video game, and non-adaptive n-back groups, with 3, 3, 7, and 3 dropouts respectively.

“In our final sample, 34 participants were assigned to the spatial n-back training with three participants dropping out.Thirty-five participants were assigned to the verbal complex span training with three dropouts, 33 were assigned to the video game training with seven dropouts, and 30 were assigned to the non-adaptive n-back group with three dropouts.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Participants who reported playing real-time strategy games weekly were reassigned to other groups to prevent skilled players from being in the RTS condition.

“If an individual indicated that he or she had played RTS games on a weekly basis, that participant was reassigned to one of the three other groups.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Studer-Luethi et al. (2012) reported weak correlations between neuroticism and training performance (r = -.24) and conscientiousness and training performance (r = .28).

“Studer-Luethi et al. (2012) reported weak correlations between neuroticism and training performance (r = -.24) and conscientiousness and training performance (r =. 28).”

• cites Studer-Luethi et al. (2012) sourceNot checked yet

  Not checked in the free run.

May combine two claims.

Recent research suggests neither length nor number of training sessions nor an adaptive element is necessary for transfer.

“This is consistent with very recent research suggesting that neither the length nor number of training sessions (Au et al., 2015) nor an adaptive element to training is necessary for transfer (Karbach & Verhaeghen, 2014;Vartanian et al., 2013;von Bastian & Eschen, 2015).”

• cites Au et al., 2015Not checked yet

  Not checked in the free run.

• cites Karbach & Verhaeghen, 2014 sourceNot checked yet

  Not checked in the free run.

• cites Vartanian et al., 2013Not checked yet

  Not checked in the free run.

Classified as probably not a substantial claim.

The study used an RTS video game group as a second active control to control for engagement and expectancies for improvement.

“We also used a RTS video game group as a second active control group to control for possible differences in engagement and expectancies for improvement between our first control and our training groups.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Correlations between training improvement and transfer have been viewed by some as evidence that transfer is moderated by training gains.

“Correlations between the amount of training improvement and the amount of transfer observed have been viewed by some researchers as evidence that transfer is moderated by training gains (Jaeggi et al., 2011).”

• cites Jaeggi et al., 2011 sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

Responder analyses have been criticized as logically insufficient and unusable as support for or against training effectiveness.

“However, such responder analyses (i.e., individuals showing the most gains during training have larger transfer) have recently been criticized as logically insufficient and unusable as either support for or against the effectiveness of any particular training program (Tidwell et al., 2014).”

• cites Tidwell et al., 2014Not checked yet

  Not checked in the free run.

May combine two claims.

Glass, Maddox, and Love (2013) reported improvements in cognitive flexibility in a Starcraft group after 40 h of managed play emphasizing resource awareness and rapid switching, while a second group playing the same 40 h without those demands showed no improvement.

“Glass, Maddox, and Love (2013) reported improvements in tasks measuring cognitive flexibility in a group trained on Starcraft, but only after 40 h of carefully managed play emphasizing maintained awareness of resources and rapid switching.A second group played Starcraft for the same 40 h, but without these specific demands, and showed no improvement.”

• cites Glass, Maddox, and Love (2013) sourceNot checked yet

  Not checked in the free run.

May combine two claims.

Sprenger et al. (2013) did not find transfer to Stroop or antisaccade and only reported transfer to a flanker task when training included a flanker-like task.

“Sprenger et al. (2013) also did not find transfer to Stroop or antisaccade and only reported evidence for transfer to a flanker task when the training regimen included a flanker-like task.”

• cites Sprenger et al. (2013)Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

Personality differences have been raised as a factor affecting transfer of training.

“Personality differences have also been raised as a factor that affects transfer of training (Jaeggi et al., 2014).”

• cites Jaeggi et al., 2014 sourceNot checked yet

  Not checked in the free run.

May combine two claims.

Improvements in inhibitory processing could fit the n-back data, but there was no evidence for transferable improvement to the Simon task or ANT.

“In terms of our study, improvements in inhibitory processing could fit our n-back data though we did not find any evidence for transferable improvement to the Simon task or the ANT.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Other research has reported no evidence of transfer of training to short-term memory or working-memory measures.

“Other research has also reported no evidence of transfer of training to SHORT-TERM MEMORY or working-memory measures (Lilienthal, Tamez, Shelton, Myerson, & Hale, 2013;Schwab et al., 2015;Sprenger et al., 2013).”

• cites Lilienthal, Tamez, Shelton, Myerson, & Hale, 2013 sourceNot checked yet

  Not checked in the free run.

• cites Schwab et al., 2015Not checked yet

  Not checked in the free run.

• cites Sprenger et al., 2013Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

One possible reason for the lack of far transfer to reasoning measures is that spatial n-back training improvements were smaller than earlier reports.

“One possible reason for the lack of far transfer to the reasoning measures in our study is that the training improvements seen in our spatial n-back group were smaller than earlier reports (e.g., Jaeggi et al., 2010).”

• cites Jaeggi et al., 2010Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

Boot et al. found no transferable improvement after 21 h of training.

“even though prior studies do not support that interpretation as Boot et al. found no transferable improvement after 21 h of training”

• cites Boot et al.Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

Individuals who volunteer and persist in a multiweek training intervention with little financial incentive may differ systematically in personality or cognitive abilities from those who participate for compensation.

“It may be that individuals who volunteer and persist in a multiweek training intervention with little to no financial incentive may be systematically different in either personality or cognitive abilities than those who participate for compensation.”

May combine two claims.

The study did not systematically manipulate lure types in the n-back paradigms, so it could not test whether n-back groups were less susceptible to near lures.

“Unfortunately, we did not systematically manipulate the type of lures in our n-back paradigms to test whether our n-back groups would be less susceptible to near lures.”

May combine two claims.

Training on the n-back does not appear to increase working-memory capacity, and the observed transfer did not appear related to self-reported strategies.

“Therefore, training on the n-back would not seem to increase the capacity of working memory and the transfer we observed did not appear to be related to the self-reported strategies used during training.”

Classified as probably not a substantial claim.

Video game training would not control for task-specific expectancies for improvement.

“Additionally, video game training would not control for task-specific expectancies for improvement such as those possibly seen in our non-adaptive n-back control group.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

The similar n-back transfer improvement raises two possibilities: both n-back groups improved on a common cognitive process, or task-structure similarity increased confidence or expectation.

“which raises two possibilities: (1) participants in both n-back groups improved on a cognitive process common to n-back tasks, or (2) that the similarity in task structures led to an increased confidence or expectation that they could perform well on the task.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Jaeggi et al. (2014) found volunteers self-reported more cognitive failures and described them as having a self-perceived deficit that may have influenced their interest in improving memory.

“Jaeggi et al. (2014) found that volunteers self-reported more cognitive failures and described these participants as having Bsome self-perceived deficit that many have influenced their interest in improving their memory and cognitive performance in the first place^ (Jaeggi et al., 2014).”

• cites Jaeggi et al. (2014) sourceNot checked yet

  Not checked in the free run.

May combine two claims.

The study used an active control group similar to previous studies, intended to control for basic task structure and materials while omitting the adaptive component believed important for transfer.

“We used an active control group similar to those used in previous studies intended to control for the basic task structure and materials of one of our working-memory training groups while omitting the adaptive component of training that was believed to be important for transfer (Klingberg et al., 2005).”

• cites Klingberg et al., 2005Not checked yet

  Not checked in the free run.

Classified as probably not a substantial claim. May combine two claims.

The first control group can be criticized as not a control but a training group based on similar transfer, and the second control could be criticized as possibly more of a training group.

“While our first control group can be criticized as not a Bcontrol^group, but rather a training group based on the similar transfer seen, our second control group could also be criticized as possibly being more of a training group”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

The study did not measure initial motivation, reasons for participating, or changes in challenge or interest across training.

“However, we did not measure how motivated our participants were at the beginning of the study, their reasons for participating in the study, or whether the degree of challenge or interest they had in their training task changed across the training.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Jaeggi et al. (2014) proposed that whether studies find transfer to fluid intelligence may depend on intrinsic motivation and that paying participants may undermine intrinsic motivation.

“Jaeggi et al. (2014) proposed that one explanation for why some studies find transfer to fluid intelligence and others do not may depend on how intrinsically motivated the participants are and that paying participants may undermine intrinsic motivation.”

• cites Jaeggi et al. (2014) sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

Other research groups have suggested that inhibition of irrelevant information in working memory may be the locus of n-back training.

“Other research groups have also suggested that the ability to inhibit irrelevant information in working memory may be the locus of n-back training (Jaeggi et al., 2014;Schwarb et al., 2015).”

• cites Jaeggi et al., 2014 sourceNot checked yet

  Not checked in the free run.

• cites Schwarb et al., 2015 sourceNot checked yet

  Not checked in the free run.

May combine two claims.

Single-domain span training is unlikely to yield transfer, and broader transfer in other studies may rely on training with multiple complex span tasks.

“Therefore, it appears that single domain training using a span task is unlikely to yield any transfer, and the broader transfer seen in other studies may rely on training with multiple complex span tasks which allows participants to attend to the shared task structure and learn to shift between task demands more efficiently.”

May combine two claims.

In our study, working-memory capacity was not increased directly, nor was working-memory efficiency improved through developing and practicing various strategies.

“In our study, it would not appear that working-memory capacity was increased directly nor was the efficiency of working memory improved through the development and practice of different strategies.”