Manfred's Column, May 2005: The focus of attention and working memory


May 2005

The focus of attention and working memory – or, why the use of cellular phones while driving should be banned

Imagine you are traversing city streets. While driving, you need to focus on important information, e.g. a red light or pedestrians walking into the street, while ignoring irrelevant information, such as the type of tree lining the road. To achieve such a feat, selective attention is required. Now imagine that, while driving, you are constantly going over the street name and number of your destination in your mind. This process requires working memory, which is where the brain temporarily stores information used in reasoning and planning. Thus, in order to get safely to your goal, you need two higher cognitive capabilities, selective attention and working memory.

These two functions are among the most frequently examined higher cognitive capabilities in humans. Working memory refers to the process of actively retaining relevant information for brief periods of time. Thus working memory keeps information “online” and thereby influences what we are doing (2). For example, when we look up a phone number, we store it in working memory and thus keep it “online” for dialling. In addition, working memory is necessary for the execution of planned actions and for the pursuit of goals. In order to do this, working memory needs to block input that collides with, interferes with or simply disturbs the goal-directed planned actions. For example, my working memory allows me to finish the article, the paragraph or at least the sentence, before I stand up to get some coffee. Without working memory, the slightest appetite for coffee would initiate the corresponding action. The neurobiological substrate of working memory lies within the frontal cortex, as single-cell recordings in monkeys or neuroimaging studies in humans have shown. Within the frontal cortex, the dorsolateral prefrontal cortex (DLPFC) is the cortical area, where the functions of working memory predominantly reside.

Attention processes allow us to direct attention to certain aspects in the environment with special care. An impressive example of the capacity to direct attention is the “cocktail party effect”: at a party, in the chaos of noise and voices, we are able to focus and concentrate on voices further away and thus follow a conversation, despite the other acoustic signals that continuously reach our ear. We are also especially good at selectively directing our attention to one or another aspect of an object, such as its colour, shape or movement. The neurobiological substrate of selective attention is encoded in the brain’s capacity to direct information processing capacity into regions where it is most needed. For example, when we concentrate on the colour of an object, heightened activity can be measured in the areas of the cortex that are involved in the processing of colours.

Nowadays, there is a large body of research on working memory and attention. However, these two cognitive capabilities have mostly been regarded separately, even though knowledge about the interaction of these two functions could be of high practical relevance, as the abovementioned example illustrates.

One of the first studies investigating the relationship between attention and working memory examined how working memory could modulate attention and affect the processing of relevant and irrelevant visual stimuli (1). Therefore, subjects had to accomplish two unrelated tasks: a task that required selective attention and a working memory task. In the selective attention task, the written names of famous politicians or pop stars were presented and subjects were asked to classify the names into the categories of politicians and pop stars. As each name was flashed on the screen (=relevant stimulus), the face of a politician, a pop star or an anonymous face was also presented (=irrelevant stimulus). This face could either match the written name (=congruent condition) or not (=incongruent condition). For example, subjects saw Mick Jagger’s face together with his written name or they were presented with Bill Clinton’s face together with the name David Bowie written on the screen. In the first case, Mick Jagger’s face did not disturb the process of classifying the written name as a pop star. In the second example however, the subjects had to ignore the face of Bill Clinton in order to correctly classify the written name David Bowie as a pop star. For assessment of distractor face processing, reaction times between the congruent and the incongruent condition were compared.

This attention task was combined with a working memory task using digit order: subjects were asked to memorise a string of 5 digits before beginning the attention task and to then recall the string of digits after completing the attention task. There were two conditions for the attention task, one with a low and one with a high load for the working memory: when digits were presented in the fixed order 0 1 2 3 4, memorising was apparently easy, i.e. led to a low load for the working memory. However, when digits were presented in a random order, e.g. 0 3 1 2 4, constant internal rehearsing of the digits was required (as in the phone number example given earlier), which put a high load on the working memory.

The tasks were performed in the following order: first, the digits of the working memory task were presented and subjects were asked to rehearse them. Second, subjects were asked to classify the pop stars or politicians in the attention task. Third, subjects were asked to recall the digits of the working memory task in a selective attention experiment. After a 500 milliseconds (ms) fixation across a display, subjects were presented the working memory task, which consisted of the memorisation of 5 digits. The task put a high load on the working memory, i.e.  5 digits were presented in random order for 1.5 seconds (0 3 1 2 4). For a low working memory load, the digits were always in the following order 0 1 2 3 4 5. After the working memory task, a fixation display was presented for 850 ms, followed by two, three or four attention displays. These displays consisted of the names of politicians or pop stars, which the subjects had to classify. Each name appeared over the face of a politician or a pop star, which was either congruent with the name or not. Each attention display was presented for 500 ms and was followed by a 1250 ms blank response interval in order to allow subjects enough time to classify the name. After the final attention display, a working memory probe was presented. Participants were requested to report the digit that had followed this probe in the working memory task. In this example they would have had to press “4” when presented with the probe “2”, since “4” had followed the digit “2” presented in the working memory task.

The results can be summarised as follows: when working memory was heavily taxed (random string of digits), the dorsolateral prefrontal cortex (DLPFC) showed a stronger activation compared to when working memory was lowly taxed. This finding shows that the memory load determined the activation of brain areas associated with working memory, such as the DLPFC, and clearly demonstrates that the experimental variation in load for the working memory was successful. But above all, the hypothesised relationship between working memory and attention could be proven: neuronal activity relating to the presence vs absence of distractor faces in the selective attention task was modulated by working memory load, i.e. distractor faces caused greater activation in inferior temporal areas, which are known to be selective for face processing, when working memory was highly loaded, compared to when it was not. This means that when the brain was thinking hard, it spent more effort processing irrelevant visual information in order to maintain stimulus priorities. In addition, reaction times for classifying the names were longer if a distracting face was presented and working memory was highly taxed. For example, the face of Bill Clinton distracted from the classification of the name David Bowie. This interference was again greater under high working memory load than low working memory load. Thus, the ability to act upon relevant information and ignore irrelevant distractors depends on the availability of working memory.

Taken together, the results of this neuroimaging study prove that working memory and attention interact: if working memory is highly loaded, distracting stimuli have a stronger activation on the cortical structures that process them. These results show that working memory plays a major role in the control of selective attention.

This finding could have practical implications. For example, it provides a strong argument against the use of cellular phones while driving. Until now, safety measures have mostly been centred on the use of headsets or speakerphones for drivers. But the findings of the described study suggest that the availability of one’s hands may only contribute a small part to safe driving: it could well be that accidents involving cellular phone usage may not be caused because of one hand holding the phone, but because working memory is occupied by the conversation over the cellular phone (3). Any phone conversation will tax working memory and may thus cause a driver to be more distracted by irrelevant sights on the road and could thereby inhibit the processing of important information, such as a pedestrian or a red light. But although this is only one of the implications that the findings of this study might have, it already shows how neuroscience could make important contributions to a more sensible and safe daily life.

(translation of Spitzer M. (2002). “Arbeitsgedächtnis und Aufmerksamkeitsfokussierung”, by, in Spitzer M. “Schokolade im Gehirn: und weitere Geschichten aus der Nervenheilkunde”, Stuttgart-New York: Schattauer, 2002, pp.87-90)


1. De Fockert JW, Rees G, Frith CD, Lavie N (2001). The role of working memory in visual selective attention. Science 291: 1803-6
2. Spitzer M. (2001). Behalten, Auswählen und Fehleraufspüren. Dissoziierbare Funktionen im Frontalhirn. In: Sptizer M. Ketchup und das kollektive Ungewusste. Stuttgart-New York: Schattauer. 95-101.
3. Wickelgren I (2001). Working memory helps the mind to focus. Science 291:1684-5