A formal model of capacity limits in working memory
- A mathematical model of working-memory capacity limits is proposed on the key assumption of mutual interference between items in working memory. Interference is assumed to arise from overwriting of features shared by these items. The model was fit to time-accuracy data of memory-updating tasks from four experiments using nonlinear mixed effect (NLME) models as a framework. The model gave a good account of the data from a numerical and a spatial task version. The performance pattern in a combination of numerical and spatial updating could be explained by variations in the interference parameter: assuming less feature overlap between contents from different domains than between contents from the same domain, the model can account for double dissociations of content domains in dual-task experiments. Experiment 3 extended this idea to similarity within the verbal domain. The decline of memory accuracy with increasing memory load was steeper with phonologically similar than with dissimilar material, although processing speed was faster forA mathematical model of working-memory capacity limits is proposed on the key assumption of mutual interference between items in working memory. Interference is assumed to arise from overwriting of features shared by these items. The model was fit to time-accuracy data of memory-updating tasks from four experiments using nonlinear mixed effect (NLME) models as a framework. The model gave a good account of the data from a numerical and a spatial task version. The performance pattern in a combination of numerical and spatial updating could be explained by variations in the interference parameter: assuming less feature overlap between contents from different domains than between contents from the same domain, the model can account for double dissociations of content domains in dual-task experiments. Experiment 3 extended this idea to similarity within the verbal domain. The decline of memory accuracy with increasing memory load was steeper with phonologically similar than with dissimilar material, although processing speed was faster for the similar material. The model captured the similarity effects with a higher estimated interference parameter for the similar than for the dissimilar condition. The results are difficult to explain with alternative models, in particular models incorporating time-based decay and models assuming limited resource pools.…
Verfasserangaben: | Klaus OberauerORCiDGND, Reinhold KlieglORCiDGND |
---|---|
URL: | http://www.sciencedirect.com/science/journal/0749596X |
DOI: | https://doi.org/10.1016/j.jml.2006.08.009 |
ISSN: | 0749-596X |
Titel des übergeordneten Werks (Deutsch): | Journal of Memory and Language |
Verlag: | Elsevier |
Verlagsort: | Amsterdam |
Publikationstyp: | Wissenschaftlicher Artikel |
Sprache: | Englisch |
Jahr der Erstveröffentlichung: | 2006 |
Erscheinungsjahr: | 2006 |
Datum der Freischaltung: | 25.03.2017 |
Freies Schlagwort / Tag: | capacity; interference; mathematical model; non-linear mixed effects; working memory |
Band: | 55 |
Ausgabe: | 4 |
Erste Seite: | 601 |
Letzte Seite: | 626 |
Organisationseinheiten: | Humanwissenschaftliche Fakultät / Strukturbereich Kognitionswissenschaften / Department Psychologie |
DDC-Klassifikation: | 1 Philosophie und Psychologie / 15 Psychologie / 150 Psychologie |
Peer Review: | Referiert |
Publikationsweg: | Open Access |
Name der Einrichtung zum Zeitpunkt der Publikation: | Humanwissenschaftliche Fakultät / Institut für Psychologie |