Example of refractory period in Psychology

Sequential Presentation Dual-Task Methodology

Another dual-task methodology for assessing capacity is the sequential presentation method. The best-known and most widely used variant is the psychological refractory period (PRP) paradigm. In the PRP paradigm, two tasks (T1 and T2) are presented in close succession with a variable stimulus onset asynchrony (SOA). Participants are instructed to make speeded responses to T1 and T2 (denoted R1 and R2, respectively), and the response time (RT) for these two tasks is referred to as RT1 and RT2. Typically, RT1 is unaffected by SOA. A ubiquitous finding, however, is that RT2 increases as SOA decreases (i.e., as temporal overlap between tasks is increased). This increase in RT2 has been termed the PRP effect.

PRP findings generally support the central bottleneck model, shown in Figure 1. This model hypothesizes that peripheral processes (encoding and response execution) can proceed in parallel for T1 and T2, but central processes (e.g., response selection) cannot. Thus, at short SOAs, T2 central processes must wait for T1 central processes to finish. This waiting produces the PRP effect on T2.

Sequential presentation dual-task methods (such as the PRP paradigm) address the two conceptual complications mentioned in the previous section on simultaneous presentation dual-task methods. First, sequential presentation (typically with same task order on every trial) tends to encourage a consistent dual-task strategy. Often, interference is restricted to T2, so that there is no performance tradeoff between tasks. Second, this approach uses several different SOAs between the presentations of task 1 (primary task) and task 2 (secondary task), mixed within a single block of trials. Mixing minimizes unwanted strategic differences between conditions. Because participants must always prepare for both tasks and hold both tasks in memory, this paradigm makes it possible to isolate true dual-task costs from the costs of preparing and maintaining two task sets in working memory.

Most studies of aging and PRP effects have shown evidence of larger PRP effects for older adults than for younger adults. However, the theoretical interpretations for this interaction differ widely across studies. Allen et al. proposed an executive control deficit that decreases the efficiency of switching between T1 and T2. Hartley and Little argued that older adults do not have an executive control deficit, but rather are less efficient in managing input interference and output interference. Finally, Glass et al. proposed that older adults exhibit poorer task coordination strategies. Although the Allen et al. and the Glass et al. viewpoints both emphasize executive control deficits, they are based upon opposing interpretations of PRP effects (the former is based upon Pashler's central bottleneck architecture and the latter is based upon Meyer and Kieras's executive process interactive control [EPIC] model). Consequently, although there is some agreement that older adults have difficulty in task coordination or time-sharing, there is no consensus as to the specific mechanisms involved.

Another major advantage of the sequential presentation method is that it provides an independent set of analytical tools to assess capacity – locus-of-slack logic. This logic allows one to determine which mental operations are capacity-limited (i.e., are subject to the processing bottleneck) and which are not. For instance, to determine whether a particular T2 process is subject to the bottleneck, one can manipulate the duration of that stage and measure the interaction with SOA. If the manipulated T2 stage occurs at or after the bottleneck, the effects should be constant across SOAs (an ‘additive’ interaction; see Figure 2A). If the manipulated T2 stage occurs before the bottleneck, the effects should decrease as SOA decreases (an ‘underadditive’ interaction, also known as ‘absorption’ of a factor effect into ‘cognitive slack’; see Figure 2B). Thus, by examining whether a manipulation has additive or underadditive effects with SOA, one can determine whether the corresponding stage of T2 is subject to the bottleneck (capacity-limited).

Research using locus-of-slack logic with younger adults has revealed that perceptual processing (up to identification, in the case of letters) is generally not subject to the bottleneck. However, many subsequent processes are subject to the bottleneck, including response selection, memory retrieval, memory encoding, mental rotation, and lexical access.

With these findings in mind, it is interesting to ask whether the bottleneck has the same locus for younger and older adults. Given the hypothesis of reduced capacity for older adults, one might assume that they would have a longer list of bottleneck processes. In fact, the opposite appears to be true. Allen et al. reported a study in which T1 was tone discrimination and T2 was a lexical decision task (word vs. non-word). The difficulty of T2 lexical access was manipulated by choosing words with either high or low frequencies of use in the English language. Although word frequency effects on RT2 were roughly additive with SOA for younger adults, they were underadditive for older adults. These results suggest that lexical access was subject to the bottleneck for younger adults, but not for older adults. Furthermore, older adults actually showed shorter dual-task RTs than younger adults, after appropriately controlling for generalized slowing. Thus older adults actually demonstrated greater capacity than younger adults, at least with respect to the lexical access stage. One explanation for this surprising finding is that older adults have automatized lexical access due to their far greater cumulative experience with words.

An important question to address is why older adults appear to exhibit capacity decrements in simultaneous presentation dual-task paradigms yet can exhibit capacity advantages in sequential presentation dual-task paradigms. One possible explanation is that simultaneous presentation methods confound executive control effects (preparing for two tasks, deciding which to do first) and capacity effects (resource conflicts), whereas the sequential presentation methods allow one to isolate capacity effects. Perhaps much of the age deficit found in simultaneous presentation divided attention tasks is due to executive control effects rather than capacity effects. Consistent with this hypothesis, task-switching studies (see later section on executive task control) have often revealed evidence of executive control deficits in older adults.

Rules and Task Coordination

In multitasking situations, participants are asked to perform two tasks simultaneously (dual-task paradigms) or in rapid succession (as in the psychological refractory period paradigm). The inferior frontal cortex, a region of the vlPFC, shows robust activation in multitasking situations. This is the case across any different pairing of stimulus (visual or auditory) or response (manual or vocal) modality (Dux et al., 2009; Stelzel et al., 2006, 2008, 2011), suggesting that the inferior frontal cortex is uniquely associated with task order control (Stelzel et al., 2008).

Although the inferior frontal cortex is robustly activated by any pairing of stimulus and response modality, the specific details of the pairing modulate the neural response. Tasks can be characterized as using a standard pairing if they pair visual–manual or auditory–vocal dimensions or as using a nonstandard pairing if they pair visual–vocal or auditory–manual dimensions. With nonstandard pairings, there are larger behavioral interference and larger dual-task related activation in the IFJ (Stelzel et al., 2006). Thus, IFJ is sensitive to the amount of overlap or the amount of translation that needs to be accomplished when mapping a stimulus dimension onto a response dimension in multitasking situations.

Central Capacity Limits

Much work demonstrates interference – that is, shared capacity limits – across disparate modalities and tasks. This research suggests that in addition to capacity limits embedded in specific perceptual systems (e.g., visuospatial perception or audition), there also exist more centrally localized bottlenecks. In a capacity-unlimited system, or when two capacity-limited processes are independent from one another, simultaneous demands on multiple processes should not slow or otherwise interfere with each other. However, when processes share a common capacity limit, attempting to carry out both processes at once should slow or degrade one or both processes. Such a capacity-limited situation occurs during the psychological refractory period (PRP) paradigm (Pashler, 1994; Telford, 1931), in which simultaneous or temporally proximate demands on the response-selection process from two distinct stimulus–response mappings lead to slowed processing of the second-presented stimulus while response selection ‘bottlenecks’ cognition and processes the first stimulus. This response-selection bottleneck (RSB) occurs even when the stimuli and responses have nonconflicting perceptual and motor modalities (Kamienkowski, Pashler, Dehaene, & Sigman, 2011; Pashler, 1994), though this bottleneck is not immutable as it can be dramatically attenuated with extensive practice (Schumacher, Elston, & D'Esposito, 2003; Schumacher et al., 2001). The RSB is linked to overlapping neural activations across tasks and modalities (Jiang & Kanwisher, 2003a, 2003b) and may be primarily neurally instantiated in the prefrontal (Dux, Ivanoff, Asplund, & Marois, 2006; Dux et al., 2009) and parietal (Sigman & Dehaene, 2008) cortex. In particular, dorsolateral prefrontal cortex and anterior insula have been demonstrated recently to activate for response selection regardless of perceptual and motor modality (Ivanoff, Branning, & Marois, 2009). In addition, these regions do not encode the perceptual or motor modality of response selection (Tamber-Rosenau, Dux, Tombu, Asplund, & Marois, 2013), further supporting the view that these regions are involved in central processes. Finally, behavioral (Jolicoeur, 1999; Jolicoeur & Dell'Acqua, 1998) and neuroimaging (Tombu et al., 2011) work suggests that this bottleneck is not restricted to response selection but instead reflects a ‘unified’ bottleneck including other attention-demanding central processes such as encoding stimuli into WM. The same neural systems involved in the RSB have also been identified in a direct comparison of the PRP and the AB (Marti et al., 2012).

Doing Two Things at Once

In the preceding examples individuals may have had to deal with multiple stimuli, but the task in each case was in some sense unitary (count the objects, repeat back all of the letters, etc.) There has also been a great deal of research in which quite different tasks have been combined; what has been at issue is the extent to which tasks can be performed without mutual interference. The general expectation based on everyday experience is that it is difficult to do several things at once (e.g., pay attention to two simultaneous conversations). In contrast, multitasking is possible in some cases (e.g., driving while listening to the radio). Thus the research effort has largely been directed at finding out just which tasks are difficult to combine, and why. One part of this effort has been a search for special conditions in which tasks might be combined without loss, as such conditions might have implications for real-world scenarios, such as using a cell phone while driving.

Some early research with continuous tasks suggested that even fairly complex pairs of tasks that initially interfered with one another – such as typing visually presented text and shadowing (i.e., repeating back) a message played through earphones – could, after sufficient practice, be done as well together as separately. However, with complex continuous tasks it is often possible to interleave components of the tasks so as to minimize or even eliminate the amount of temporal overlap between components of the two tasks. (An extreme example: a centerfielder could probably read a book during the ‘slack time’ when he is not actively engaged in the game.) Scientists wishing to achieve a more fine-grained analysis of dual-task performance have used simplified tasks to permit a closer analysis of stimulus–response relations. Perhaps the most well-developed analyses come from the psychological refractory period paradigm. In this paradigm two stimuli are presented, S1 and S2, with a variable stimulus-onset asynchrony (SOA). Typically, study participants are instructed to respond as quickly as possible to each stimulus with the appropriate response, R1 and R2, respectively. When the two tasks are widely separated in time there should be no interference between them and we can obtain baseline reaction times (RTs). As the stimuli are presented closer together in time, the reaction time to the second stimulus (RT2) will tend to increase. Most strikingly, in some cases, as the time between stimuli becomes less than about 300 ms, RT2 increases about 1 ms for each millisecond that the SOA is decreased. It is this kind of result, suggesting a clear inability to produce R2 at the same time as R1, that led to the nomenclature of a refractory period, by analogy to the refractory period of neurons.

Another example of the difficulty of responding to two stimuli presented close together in time is known as the attentional blink (AB). In a typical AB task, a rapid series of stimuli (e.g., digits or letters) is presented at fixation, typically at a rate around 10 items per second, and either one or two targets can appear within the stream. The AB refers to a decrement in the detection or identification of the second target (T2) when it occurs soon after the presentation of the first target (T1), to which a response is required. Early models of the attentional blink assumed that T2 processing is impaired because attention is temporarily fully occupied by processing of T1.

In an instructive example, a stream of letters was presented at fixation. There were two targets. The first was a green letter; all of the other letters in the stream were black. The green letter was followed after a variable delay (the lag) by the second target, which was a ring of Gabor patches (a Gabor patch is, essentially, a small patch of parallel stripes) in the periphery surrounding one of the later letters in the stream. Participants had to name the colored letter and also indicate if all of the Gabor patches were oriented in the same direction or if one was misoriented by 90°. In a control condition, individuals could ignore the stream of letters and just indicate if the ring of Gabor patches contained an orientation oddball. Performance in the control condition was about 90% correct and did not vary with the lag between the green letter and the Gabor patches. In the experimental condition, performance was poor (about 60% correct) when the Gabor patches were simultaneous with the green target letter, and improved to nearly 90% correct when the lag between the green letter and the Gabor patches was 700 ms.

What makes these results interesting are the implications they may have for the study of visual processing. The detection of a Gabor patch differing by 90 ° in orientation from an otherwise uniformly oriented set of patches should be handled preattentively according to feature-integration theory, and, indeed, the reaction time to detect a Gabor patch misoriented by 90° is independent of the number of other patches – which is precisely the diagnostic that has been taken to indicate preattentive processing in discussions of feature-integration theory. Why then should there be a large dual-task decrement? One possible answer is that it takes a really difficult competing task, such as is presumably provided by the attentional blink paradigm, to show that orientation discrimination, while easy, is not completely attention free. However, it is also possible that the deficit is not due to T1 stressing attentional capacity for feature processing; it may be that the attentional blink reflects difficulty with one or more higher level functions, such as maintaining executive control of a task set in the face of a fast-moving sequence of targets and nontargets. The attentional demands of various kinds of perceptual discriminations remain a topic of great interest and the final word has not yet been written.

What happens during a refractory period psychology?

Why is refractory period important in psychology?

What are the 2 types of refractory periods What is the function of each?