Assorted Caveats and Outstanding Issues

An aspect of the present data which needs consideration is the difficulty of manipulating motivational levels across many individuals to a similar degree simply by changing externally presented cues. Personality together with different life experiences presumably influence individual coping and learning strategies, as well as reactions to positively or negatively motivating events. The experiments presented in this work might therefore be seen as a procedure to influence high and low levels of motivation in a rather artificial, yet controlled way. This should not discount the relevance of the obtained results as they considerably extended previous findings of basic research in this domain and at the same time will serve as a crucial starting point for future studies. It remains to be determined however whether the results obtained in the three studies that were presented in this thesis would also be obtained in everyday learning and retrieval situations outside of a controlled laboratory environment. Additionally, the present analyses apply to memory for rather simple events (words, black-and-white drawing pictures) of the kind frequently used in and typical for experimental studies. By this and the fact that the dependent and independent variables of interest were kept as simple as possible and in a relatively low, assessable number, advantage was taken of the possibility to control for potential confounding of less or uncontrollable variables that might influence an effect. The disadvantage of this is that laboratory memory studies usually differ to a large extent from everyday life situations, e.g. by the use of EEG-equipment, the requirement for participants to sit in front of a computer monitor and to respond via button press on the keyboard. Therefore, it remains to be explored whether similar reward-associated retrieval orientation effects would also be obtained in more ecologically valid learning and retrieval situations. This would permit further generalization of the experimental findings that were presented in this thesis.

A further issue in this context is that the test phases in the conducted experiments were, next to the target-material conditions always blocked by high and low reward. This design aspect was employed on the basis of the findings of Werkle-Bergner et al. (2005) which showed that ERP retrieval orientation effects disappeared under frequently changing testing conditions and high task-switching demands. However, designs that are optimal for recording separate contrasts between reward and material conditions at retrieval are not always compatible with one another. Therefore, it may be more appropriate to conduct experiments that try to use memory tests in which test items of different conditions would be interleaved within a single test block but in such a way that sequences of same items are still

retrieval orientations. Moreover, by this the possibility that some participants might have uncovered the reward blocking conditions in the test phase would be ruled out, although the outcomes of the participant’s questionnaire indicate that this was unlikely in the current design.

Furthermore, it might be that the limited amount of time to give a response in the memory exclusion task negatively influenced the recognition outcome for some individuals or some trials due to excessive time pressure. Despite the fact that limiting respondents’

answering time for example in memory tests might discount large individual differences in the time needed to give a response, the particular time parameters employed in the studies presented in this work were chosen in order to reduce uncontrolled influences of distraction, body movements or boredom. It is very likely however that if more time was given to answer, participants would have been able to correctly recognize more targets, non-targets and new items. Support for this comes from the improved memory performance that was attained in the series of the four cross-cultural, behavioral experiments (Experiments 3a-d). However, the extent to which longer response times might have led to increased memory accuracy remains ambiguous because alongside longer response time windows, these experiments also consisted of approximately 30% fewer trials at study and test compared to the experimental designs used in Experiment 1 and Experiment 2.

Whether better memory for items that were preceded by a high reward or a high punishment cue during study as observed most clearly in the behavioral experiments (Experiments 3a-d) is purely based on the fact that participants were highly motivated especially during encoding is also open to question. A probable disadvantage of the relatively long 24-hours retention interval between study and test might be that additional time was given to retrieve the newly learned information at several, uncontrolled time points. This implies that participants may have additionally strengthened the memory trace selectively for highly motivating information by the repeated search for internal or external retrieval cues in order to reactivate the trace. This possibility would extend the scope of the present research project, but it would be highly interesting to disentangle the factors that contribute most to memory enhancing retrieval processing.

In a broader sense, it might be of interest for clinical purposes to further develop the results obtained in basic research in order to increase their implications for neuropsychological research and patient studies. For example, the present findings might indicate a way to improve successful memory retrieval, i.e. in the healthy elderly that undergo normal decreases in long-term memory performance with increasing age. This could be done

perhaps by improving the way retrieval cues are processed to increase their similarity with targeted memory representation as well as to optimize retrieval cues’ dissimilarity with irrelevant and non-targeted memory representations. A possibility would be to ameliorate the fit between adopted retrieval orientations to particular retrieval goals by including motivational factors during learning and maybe by focusing attention on those features of a retrieval cue that overlap with those of a targeted memory representation. By this, processing resources might be minimized. Therefore, studies with elderly participants that attempt to integrate present and prospective findings with regard to critical factors that impact positively on learning and strategic retrieval processing, would be desirable in the future.

Finally, the findings of reward-associated ERP-retrieval orientation effects presented in this thesis imply that current concepts about the functional role of retrieval orientations should be extended to the possibility of aligning retrieval cue processing to types of memory representations which did not differ either in the material encoded at study (e.g. words versus pictures) or with regard to the task requirements at test (e.g. retrieval of ‘general’ versus

‘specific’ features of studied items) but differed solely in motivational level during learning.

This means that for optimal retrieval of information, retrieval orientations are not only adopted in order to constrain retrieval cue processing at a conceptual/semantic level but also to more contrasts of retrieval conditions in which retrieval cue processing might be constrained in different ways. The observed reward-associated retrieval orientations demonstrate that incentives to memorize an item influence not only encoding but also retrieval processes.

A general requirement in empirical research is the future replication of newly presented findings to increase their validity. Of importance here is that this requirement has to some extent already been fulfilled because the critical findings of the presented work - ERP indices of adopted, reward-associated retrieval orientations - have been successfully replicated across the two ERP-studies reported in this thesis. However, replication, especially with regard to the cross-cultural changes in sensitivity to negative learning events, would be highly desirable.

‘[…] we make search in our memory for a forgotten idea, just as we rummage our house for a lost object. In both case we visit what seems to us the probable neighborhood of that which we miss. We turn over the things under which, or within which, or alongside of which, it may possibly be; and if it lies near them, it soon comes to view (James, 1890, p. 654).’

More than a century ago, William James described with this metaphor memory retrieval as a search process, an influential description which still influences the contemporary understanding of memory retrieval. In the context of the studies presented in this work in which the main goal was to examine how motivated learning, either through positive or negative monetary incentives, might influence strategic memory retrieval processes, William James’ metaphor might be extended to the figurative example of a further search process in one’s own house but now for lost and very valuable objects. As a consequence, the search process through the house might be more efficient in time and space, but the extent to which the object is likely to be found would vary depending upon the amount of value bestowed upon the object when it was originally left. This value would then affect the retrieval operations engaged during the search for it.