You are using an old browser with security vulnerabilities and can not use the features of this website.

Here you will see how you can easily upgrade your browser.

When Similar Things Interfere – Researching Working Memory

Working memory is a buffer of the human brain. Its capacity is limited, even more so in dyslexic people. Photo: Fotolia/lassedesignen

The capacity of the human brain's working memory is limited. Photo: Fotolia/lassedesignen

Working memory is a buffer of the human brain. Its capacity is limited, even more so in dyslexic people. The psychologist Katrin Göthe experimentally analyzed what exactly limits it. How to improve the performance of working memory has yet to be researched.

“Memorize the sequence of numbers and repeat them,” says Katrin Göthe. 4, 7, 3, 1, 9. This is easy. “Now subtract 4 from the last number and replace the last numeral with the result. Can you repeat the complete sequence?”

Katrin Göthe, research scientist at the Department Cognitive Psychology, has to go into greater detail to explain her recently completed project “Process dissociations of working memory functions in cases of cognitive impairment” . The brainteaser is meant only to explain the difference between short-term and working memory. Short-term memory suffices to recollect the number series, while working memory is used in the next step, in which the brain has to store information, process additional information, and block irrelevant messages. When reading a sentence we would not be able to remember its beginning without our working memory.

Our working memory is extremely useful. “Unfortunately, its capacity is limited,” Göthe explains: “To a greater extent in some than others.” Different hypotheses address the reason for these differences. A theoretical model developed at the University of Potsdam is based on the assumption that the capacity of working memory arises from interference. This means that information processed and stored interferes with other information and this even more pronounced when the information is very similar. The model distinguishes between two interference mechanisms: feature overwriting and confusion. When we have to memorize several words and two of them begin with the same letter or have other common features, one word can partly “overwrite” the other. This means that the “shared” letter gets lost in one of the two words. The word is less well represented in our memory, and we are less likely to recollect it. If there are other similarities, e.g. a phonetic similarity, the two words may be completely confused.

Tests in adults have already corroborated this model. “In my project I examined its application to dyslexic and learning disabled children,” Göthe says.

Dyslexia is a “partial performance impairment”. Dyslexic children have normal intelligence. Research, however, has shown that their working memory functions worse. According to the model, the reason lies in its greater susceptibility to interference – a potential reason why their reading and writing skills lag behind their peers.

Furthermore, Göthe wanted to verify an observation made during previous studies, which contradicts all expectations: Dyslexic children generally perform more poorly than their non-dyslexic peers when it comes to solving tasks that are demanding for working memory, but when they have to memorize many or very similar things, i.e. working memory is even more stressed, this deficit vanishes although they actually should be performing worse.

The psychologist carried out different experiments with over 120 children of different age groups with and without dyslexia. They had to complete so-called memory updating tasks. One task tested verbal working memory. The children were shown a basket with apples, bananas, and plums on a screen and had to say how many pieces of fruit of each were in the basket after a certain number was either added or subtracted. The second task referred to the spatial-visual part of working memory: First, subjects had to memorize the position of a mouse and a cat on a grid with nine boxes and their position after the children had mentally moved them several times.

Göthe measured the proportion of correct answers in relation to the duration of presentation and compared this data with the proportion deduced from the theoretical model. The results for the verbal and spatial area show that the interference model basically also applies to children.

The results regarding the differences between the young test subjects with and without dyslexia differ though. As expected, the model established a higher rate of interference due to confusion in dyslexic children than in their peers without this disorder. For feature overwriting, it was exactly the opposite: Dyslexic children performed much better – a result shown in previous studies. “We have yet to find an explanation for this strange reversal when it comes to feature overwriting,” says Göthe. “Our current project can only replicate it in the spatial area but not in the verbal. There are doubts whether we should attach a great deal of importance to this counter-intuitive result.”

Back to the brainteaser and the explanations at the beginning of the interview: Can people with a less efficient working memory improve it by training? After all, people with a greater working memory capacity have better speaking and reading comprehension, acquire complex skills more easily, like a programming language for example, and are more successful in school. It is even related to higher intelligence.

Göthe has dealt with this before and knows that under certain circumstances, training is possible – but not by simple brain training exercises like solving a crossword puzzle or Sudoku. “Such exercises will increase your performance in solving this specific type of task,” the researcher explains. “According to our current state of knowledge, this does not even mean that you can solve other, similar tasks any better. This increased performance cannot be transferred to other cognitive abilities. In other words: Such training does not help you to better remember where you parked your car or put your keys.” In fact, it is important to do exercises that train working memory rather than short-term memory.

The psychologist Klaus Oberauer, who was involved in the development of the interference model used by Göthe for her studies, identified three functions that define working memory: simultaneous storage and processing of information, integration and correlation of information, and supervision of these processes. Oberauer and his colleagues systematically analyzed which of these three functions train working memory the most. The result shows that simultaneous processing and storage of information is the most promising candidate. “Nevertheless, we have to conduct more research before we will be able to offer an effective and reliable method for comprehensive memory training,” summarizes Göthe.

The Researcher

Dr. Katrin Göthe is a research scientist at the Department Cognitive Psychology. Her research interests are working memory, the limits of parallel cognitive processing and (mathematical ) modeling of cognitive processes. The theoretical model, on which she based her recently completed project, was developed by Prof. Reinhold Kliegl and Prof. Klaus Oberauer at the University of Potsdam.


Universität Potsdam 
Department Psychologie 
Karl-Liebknecht-Str. 24–25
14476 Potsdam
E-Mail: Katrin.Goethe@uni-potsdam.nomorespam.de

Text: Sabine Sütterlin
Translation: Susanne Voigt
Online-Editing: Silvana Seppä
Contact Us: onlineredaktion@uni-potsdam.nomorespam.de