Aarhus University Seal

Enduring experiences - interaction between episodic and semantic memory in learning from games

Project: Research

  • Department of Psychology
See relations at Aarhus University


My idea for my ph.d.-project sprang from the simple observation, that people often use mental imagery or refers to specific experiences, when they try to remember abstract knowledge.
Popular everyday examples include remembering where you read a book to think of its contents, or using a mental walk around your house to recall the number of windows (Rubin, 2006).
Funnily enough, so far pedagogical psychology has paid little attention to the role of episodic memories.
Inspired by recent advances in the study of autobiographical memory, my PhD. proposes to shed light on this interesting cognitive dynamic. Empirically, I mix behavioural measures with fMRI, to observe episodic “keys” in use. My case will be the mobile audio drama De Udvalgte (trans: The Chosen Ones). The goal is both to examine the everyday interaction between episodic and semantic memories, and to apply findings towards the design of pedagogical methods.
• When and how is conscious reconstruction of episodic scenes important to recalling semantic information?
• When and how does this influence the integration of newfound information with a person’s existing knowledge structures?
• What kinds of learning experiences elicit later conscious recall, and how can this relationship be utilized in learning designs?
My project thus bridges three disciplines: Pedagogical psychology, the (neuro)cognitive study of memory (in partnership with CFIN, attachment 3), and the design of digitally mediated learning experiences (in partnership with CFEM, attachment 4).
Theoretical background
For the purpose of this study, learning is viewed in terms of memory. The need to consider personal experiences in learning was put forward in the early 20th century (Bartlett, 1932; Dewey, 1938; Lewin, cit in Kolb, 1984), but lacked the current understanding of memory. The literature simply fails to describe what makes some learning experiences elicit conscious recall, and how that relates to learning.
Theories of experiential learning now hold that competence is constructed over time (Woolfolk et al, 2008; Kolb, 1984) and that some phases rely more on explicit rumination than others (Dreyfus & Dreyfus, 1980). By implication, the use of episodic keys may prove to be a more useful cognitive strategy at certain levels of expertise than others.
The theoretical distinction between episodic and semantic memory formats (Tulving 1983, 2002), and their interaction (Schacter, 2001; Landowsky-Brooks & Alcock, 2007) may bridge this gap. Notably, the view has emerged that autobiographical memory is based on both episodic and semantic formats (Berntsen, 2009), and that the availability of raw episodic memories is limited to a matter of hours, after which long term access becomes dependent on the memory’s place in autobiography (Conway, 2005).
In the case of De Udvalgte, I predict that the drama experience can create access-“keys” to curricular information otherwise fallen prey to the transience of memory. Episodically formatted memories already contain a great deal of information (Conway, 2001), but demonstrations of how mental reliving allows people to remind themselves of facts and details (e.g. Rubin, 2006) have largely been anecdotal.
Empirical field
De Udvalgte (for a full description, see Grønbæk et al., 2010) was designed by CFEN at Århus University to teach science to 7-9th-graders. Mobile phones guide participants through an interactive learning drama physically set in Hasle Bakker. Throughout the game, the unfolding “eco horror” story is mixed with academic challenges, which leads to an unusual cross between orienteering, labwork, game and drama. Later, the scientific reports produced during the game are used in the classroom. A very distinctive situation, compared to schemas (Bartlett, 1932) for life in the classroom.
But how does this compare to just reading a book? Is curriculum remembered better? Differently? Because consistent documentation for the pedagogical efficacy of learning games is lacking (Henriksen, 2010; Weigerif, 2007; attachment 4), my project has an opportunity to fill a persistent practical and scientific gab. The blend of pedagogical psychology and memory research has newsworthy scientific merit, and demonstrates the interplay between science, design and commercial potential at Århus University.
De Udvalgte offers a ready-made experience saturated with academic information. Many school classes have already been through the game, and many more will follow in 2011-2014. This supplies a group of participants ready for empirical studies, which can be carried out at the schools by teachers equipped with well-designed surveys. CFEM have offered themselves as partners for this research project. Together, we believe that reliable psychological measurements will supply critical documentation and future directions for digital learning designs.
Research approach
The PhD. has two layers: Behavioural studies measuring performance and memory attributes, and a functional neuroimaging paradigm designed to gain a glimpse of the key-effect in action.
Participants will be recruited from school classes in and around Århus who have played De Udvalgte. The partnership with CFEM ensures commitment from schools and dedicated teachers. If need be, adult volunteers will be recruited from undergraduate courses for fMRI. The studies will all compare the recall for “distinctive” (i.e. information acquired through De Udvalgte) to “classroom” controls (i.e. reading).
The quantifiable measures will be analyzed with the ultimate goal of gauging curricular retention in relation to the use of episodic keys. School classes are not homogeneous, so in addition to the general methodological trouble in working with 12-13 year-olds, results need to be controlled for age, general academic standpoint, and other individual differences. An existing inventory such as WJ III (Schrank et al, 2002) will fit the general requirements for the study nicely, but some scales must obviously be designed to purpose.
Data will be analysed using either ANOVA-based or correlational techniques depending on the final pool. The individual approaches will thus be used to qualify each others’ findings.
1.1 Survey
To form a basic picture, students will be tested for their retention for De Udvalgte’s curriculum in line with national school tests (evaluering.uvm.dk), and on self-report scales for use of episodic keys, vividness of recall, and certainty. This can be achieved with a standardized questionnaire, but some answers (e.g. free text) may need to be translated into quantifiable measures, based on discourse mapping (Clarke, 2005). A pilot run is needed to determine whether online surveys can be used, or if a teacher or myself should administer the test verbally.
1.2 Discourse
To document learners’ subjective experience and meaning-making, a second approach analyzes how a smaller sample of learners frame and retell their knowledge, using discourse mapping and response-time measurements on taped 20 minute sessions (Clarke, 2005). If deemed interesting, it can be done in groups of 4 or 5, modelled after published studies of group reminiscing (e.g. Middleton & Brown, 2002). Participants will be cued in different manners through a semi-structured interview utilizing direct prompts for scientific information related to De Udvalgte and asking students about specific segments of the drama.
1.3. Longitudinal study
Studies 1.1 and 1.2 supply potential methods and a baseline for longitudinal measures. Recall could thus be reliably studied across time, by replicating study 1.1 with the same students every 6 months for the duration of the PhD. project, and possibly longer if CFEM should chose to carry on the documentation process. As students age, questionnaires could be sent by email.
2. Functional neuroimaging
fMRI Neuroimaging offers a second experimental layer. This setup can grant us a real-time glimpse of participants rummaging through episodically formatted memory, to recall semantic information. This approach has been validated by the cross-disciplinary Centre for Functional and Integrative Neuroimaging (see attachment 3). The setup will be modelled after published fMRI studies of memory recall (e.g. Rubin, 2006; Addis et al., 2006). To ensure that the episodic “key” is likely to be used, participants will be cued with direct references to a prior trip to Hasle. After establishing a baseline, 30 participants will be asked to recount the curriculum from De Udvalgte in the scanner. Predetermined regions of interest chosen from networks commonly associated with episodic scene construction and semantic processing (e.g. Hassabis & Maguire, 2007; Addis et al, 2006 and Martin & Simmons, 2008), will be monitored for relative changes in cerebral bloodflow. A post-scan interview will determine to what extent each participant used episodic keys to access the information from memory. The relative magnitude of activation across the recall processes will be analyzed using SPM and MatLab.
Effective start/end date01/02/201101/02/2014

ID: 128938075