We, at the Computational Memory Lab study human verbal memory behaviour and its basis in cognitive and neural processes. There are several approaches we take towards our research, including mathematical modeling, measures of behaviour in the cognitive psychology tradition, and measures of brain activity using electroencephalography (event-related potentials and oscillations) and functional magnetic resonance imaging. We like to address (and even sometimes, resurrect) classic studies, and re-examine them in light of up-to-date knowledge and novel data-analysis techniques.

For example, it is common for people to have to keep separate two associations that share a common item (e.g., first KIM-KRIS go together, then KIM is repaired with KANYE). In classic cognitive psychology work that peaked in the 1960s-80s, the thinking was that such associations are remembered independently of one another. This is counterintuitive to most people. With some twists on the data-analyses, amounting to include the appropriate control measurements, the lab has showed for the first time, definitively, that these kinds of associations do in fact compete directly with one another. The next goal is to pinpoint exactly what (cognitively as well as brain-activity wise) produces this competition and how people manage to overcome this challenge to memory, which they clearly do (such as bilinguals, who have two words for most objects).

At the heart of our Lab's research, there are mathematical models of memory behaviour. The behavioural, modeling and brain-activity work is designed to test model assumptions and speak to major, ongoing debates. For example, some models assume participants learn an ordered list by linking each list-item to another list-item ("associative chaining"); others assume participants assign each item a position or order-value (“positional coding”). Recently, they began to investigate memory strategies that real people voluntary choose to use. Our procedures enable a higher level of experimental control in strategy research. Some such strategies bear an uncanny resemblance to popular formal models (e.g., the famous Method of Loci and the Peg List method both resemble positional coding, whereas the Link method is a very literal analogue of an associative chain). Computational modellers and strategy researchers appear unaware of one another’s work. Current goals are to use instructed (compliance-verified!) mnemonic strategies to inform the debate between computational models, and to use ideas from those models to understand how those same strategies actually work to support memory effectively.

The Lab has a particular interest in oscillations, rhythmic brain-activity, which are increasingly thought to play important roles in cognition, including memory. Dr Jeremy Caplan co-devised, tested and developed an improved means of quantifying oscillations (rhythmic brain activity), now known as Better OSCillation detection (BOSC; library available in supplementary materials of Whitten et al., 2011, NeuroImage), and is starting to be adopted by other labs. The BOSC method is more selective for rhythmic activity than other approaches, and produces more consistent results across recording sites, tasks and even species.

The lab is part of the Alberta Cognitive Neuroscience Group and the Neuroscience and Mental Health Institute. Our EEG equipment is funded by the Canada Foundation for Innovation and the Natural Sciences and Engineering Research Council of Canada and we receive operating funding from the Natural Sciences and Engineering Research Council of Canada.