Visual Patterns Test/Matrix Patterns Task

The Visual Patterns Test (Della Sala et al., 1997, 1999) is derived from the visual matrix memory protocols of Phillips and colleagues (Phillips & Christie, 1977; Phillips, 1974).

Please see figure 2.2. for an example of this task. In the recognition format, the participant has to determine whether a currently presented matrix pattern, made up of a novel configuration of black and white cells, is the same or different than one previously shown.

More recent versions (VPT) of the task enable participants to recall the matrix pattern, cell by cell.

Logie and Pearson (1997) successfully used both the Visual Patterns Test recall and recognition versions in their investigation to show the separation of visual and spatial working memory components. The VPT was used alongside the Corsi Blocks Task in a developmental study and it was discovered that the systems which underpin the two separate tasks actually develop at different rates. This pattern of developmental fractionation was taken as evidence for discrete visual and spatial processes in working memory and appeared congruent with the constructs of the Visual Cache and Inner Scribe (Logie, 1995, 2011).

31 Memory for the matrix patterns was seen to be increasingly higher across the age ranges (5-6, 8-9 and 11-12 years) suggesting that the development of the visual working memory component (Visual Cache) was at a greater rate than the suggested spatial component (Inner Scribe), however Hamilton (2013) has proposed that this pattern of developmental fractionation may disappear when the VPT and Corsi Blocks scores are standardised. As with the Corsi Blocks task, there has been increasing evidence to suggest that executive resources may contribute to VPT task performance (Brown & Wesley, 2013; Hamilton et al., 2003;

Rudkin et al., 2007).

The duration of the encoding image within a memory task is one that needs to be considered with caution. Originally, Phillips (1974) suggested that an encoding time of 100 milliseconds was enough to encode the information successfully without the use of any verbal strategies or forgetting. This meant that participants had the time to encode the visual information without attempting to use any form of verbal representation. Similarly, with the use of a change detection protocol, Luck and Vogel (1997) contrasted both 100 milliseconds and 500 millisecond encoding intervals in their experiments, proposing no difference in recall accuracy of the retrieval items. These researchers also suggested that a 500 milliseconds encoding duration may not be sufficiently long enough to use any verbal coding or strategy use. However, studies investigating effects of visual working memory capacity have suggested that proactive interference effects can occur when participants are being presented with a series of trials (Hartshone, 2008). Lin and Luck (2012) on the other hand, used a five item colour change detection task and demonstrated no such effects with regards to a 100 millisecond maintenance interval.

In this context, the VPT was originally thought to be a purely visual memory task, in terms of having a visual format of representation. However, subsequent research has employed sustainability longer encoding times which have afforded opportunities for executive

32 resources and attentional control processes to semantically elaborate the visual pattern, despite their novel configuration (Brown & Wesley, 2013; Logie & Pearson, 1997; Rudkin et al., 2007). Rudkin et al. had used a 2 second encoding interval and proposed that this duration was enough to allow the use of executive attentional resources and long term memory to occur. Hartshone (2008) also proposed proactive interference effects with regards to colour change detection task, encouraging the current thesis to use a shorter encoding duration (of 500 milliseconds) to eliminate the possibility of proactive interference effects.

The contribution of executive resources was evident in the research findings of Rudkin et al.

(2007, study 1). The contribution of executive attentional resources to the VTP task performance has been extensively researched by Brown and colleagues (Brown, Forbes &

McConnell, 2006; Brown & Wesley, 2013).

In the recent work, Brown and Wesley (2013) provided strong evidence that the VPT representation within working memory may not be purely visual as first thought. The work of Brown et al. (2006) identified that participants could verbally label matrix patterns and consequently they developed high and low verbal coding of the VPT stimuli. The High Coding stimulus pattern set would mean that participants could more readily create a verbal label for the pattern and the Low Coding category meant that participants had more difficulty in creating a rich form of verbal label. The high verbalisable sets of images lead to higher VPT task performance. Brown and Wesley (2013) made use of an executively demanding secondary task, random interval tapping, in order to remove the performance advantage for the high verbal VPT.

Results from the study demonstrated that the executive attentional interference removed the advantage of the high verbal coding category. The findings of Brown and Wesley (2013) suggested that this proposed visual working memory task may indeed use verbal semantics and recruit executive resources as part of the maintenance process.

33 These conclusions can be accommodated within the Multi-Component Models of working memory from Baddeley (1974, 2000, 2012) and Logie (2011) who identified the functional interaction of each component with the use of an Episodic Buffer and executive resources.

However, it should be noted that in these models, the VPT representation has been conventionally identified more with the VSSP and Visual Cache respectively, rather than a representation that emerges from slave and executive resource interaction.

Figure 2.2. An example of the Visual Patters Test used by Brown and Wesley (2013) A and B represent patterns from level 8. C and D represent patterns from level 11. The memory test

method is the recall version rather than the recognition version.