Shifting between cognitive tasks comes at a cost, which is usually reflected in slower response times and increased error rates when trials involve a task switch (for a review, see Grange & Houghton, 2014; Rogers & Monsell, 1995). The task-switching process allows the examination of different types of cognitive costs. The first type can be measured by comparing task performance between mixed-task blocks and single-task blocks (Kray & Lindenberger, 2000). Possible labels are global costs, set selection costs, or mixing costs (Kiesel et al., 2010; Mayr, 2001; Reimers & Maylor, 2005). However, it should be noted that there is another definition for mixing costs, when referred to as being non-switch specific (Huff, Balota, Minear, Aschenbrenner, & Duchek, 2015; Marí-Beffa & Kirkham, 2014). Non-switch specific costs are defined as the difference in task performance between non-switch trials (repetition of task A or B without a switch between the tasks) within the mixed-task blocks and single-trials within the single-task blocks (Cragg & Chevalier, 2012; Karayanidis et al., 2011). The latter form of mixing costs was investigated in the present study and is referred to as general switch costs. The underlying cognitive process signified by the general switch costs is the sustained maintenance of the multiple task sets in working memory as well as the selection of the relevant task set over the alternative one (Cragg & Chevalier, 2012; Marí-Beffa & Kirkham, 2014; Reimers & Maylor, 2005). Grange and Houghton (2014) pointed out that task repetitions within mixed-task blocks take more time than pure repetitions within the single- task blocks, even though both trial types are, strictly speaking, repetition trials.
Switch costs has been proved to be a robust phenomenon found across different task-switching paradigms (alternating runs paradigm, task-cuing paradigm, intermittent instructions paradigm, voluntary task selection paradigm). Here, I shall briefly introduce two competing models that initially tried to explain the source of the observed task- switching costs in task-switching paradigms. The first model was originally proposed by Rogers and Monsell (1995), while the second by Allport (1994). According to the first model, there are two components that contribute to the existence of switch costs. The first is an endogenous component related with top-down reconfiguration of task-sets prior to stimulus presentation. It has been observed that if the cue-stimulus interval in the task- cuing paradigm (Meiran, 1996) or the response-stimulus interval in alternating runs paradigm (Rogers & Monsell, 1995) is increased, then this prolonged period to prepare for the next trial results in reduced switch costs. However, even with very long preparation time switch costs do not disappear entirely. These observations led to the proposal that there is a further task-set reconfiguration process, dependent on the actual presentation of the stimuli, termed as the exogenous component of task-set reconfiguration. According to the second model proposed by Allport (1994), switch costs are explained as a result of interference between the competing task sets (for a review of findings, see Kiesel et al., 2010; p. 861-868.). This interference can arise either as a result of persisting activation of currently irrelevant task-sets, or as a result of persisting inhibition of the currently relevant task-set.
The task-switching paradigm does not only allow us to study the influence of postural control demands on cognitive processing with regard to cognitive flexibility (i.e., switch costs) and maintenance of concurrent task sets (i.e., mixing costs), but additionally provides the possibility to determine the influence of postural control demands on between-task interference, measured as the between-task congruency effect (Meiran, 2005). In order for congruency effects to occur, bivalent stimuli that can be applied to either task are necessary. If a stimulus requires the same response for both tasks (e.g., a right keypress), it is congruent, while if it requires different responses (e.g., a right keypress in one and a left keypress in the other task), it is incongruent. The congruency effect denotes the finding that usually participants respond faster to congruent compared to incongruent stimuli (Koch and Allport, 2006). Several authors (e.g., Rosenbaum et al., 1986) have suggested a key role of motor programming in accounting for congruency effects. Accordingly, performance benefits in congruent compared to incongruent trials are due to the maintenance of the appropriate motor programming parameter in memory from trial to trial (e.g., if the same finger is used to respond), while motor parameters must be re-programmed in incongruent trials thereby increasing reaction times. Please note, that we take a broader perspective by considering parameters specifying motor programs as part of the task. In order to explain the congruency effect, it has been argued that incongruent stimuli activate both the response according to the currently relevant task rules (i.e., the relevant task set) and the response according to the currently irrelevant task rules (i.e., the irrelevant task set) of the competing task (Kiesel et al., 2010). The congruency effect reflects the inability to shield the currently relevant task set from the currently irrelevant task set (see e.g., Dreisbach
Given that the main focus of the present study was on the transfer of task-switching training, training effects will only be discussed briefly. The first finding that all age groups and all training groups were able to reduce specific switch costs as a function of training (see Figure 13) is consistent with previous results regarding younger and older adults, showing that both age groups reduced the specific costs to a similar extent (Bherer et al., 2005; Kramer, Hahn, et al., 1999; Kray & Lindenberger, 2000; Minear et al., 2002). Moreover, the results with respect to children extend prior evidence by showing that not only general switch costs (Cepeda et al., 2001; Eber & Kray, in prep.), but also specific switch costs can be reduced after intensive training. These findings are supported by the corresponding effect sizes, ranging between .85 and .98 for the group that was only trained in taskswitching (group 2). However, the fact that reliable specific switch costs were still found after extensive training (i.e., four sessions with a total of 1768 trials) suggests that specific switch costs are a relatively robust phenomenon, at least in the type of switchingtask applied in this study, that is, an alternating runs paradigm without external task cues. Nevertheless, the training benefits found in all three age groups indicate that even in children and older adults, cognitive plasticity seems to be considerable, arguing against the often reported observation of reduced training benefits for older compared to younger adults (e.g., Baltes & Kliegl, 1992; Lindenberger & Baltes, 1995; but see Kramer & Willis, 2002).
The task-switching paradigm does not only allow us to study the influence of postural control demands on cognitive processing with regard to cognitive flexibility (i.e., switch costs) and maintenance of concurrent task sets (i.e., mixing costs), but additionally provides the possibility to determine the influence of postural control demands on between-task interference, measured as the between-task congruency effect ( Meiran, 2005 ). In order for congruency effects to occur, bivalent stimuli that can be applied to either task are necessary. If a stimulus requires the same response for both tasks (e.g., a right keypress), it is congruent, while if it requires different responses (e.g., a right keypress in one and a left keypress in the other task), it is incongruent. The congruency effect denotes the finding that usually participants respond faster to congruent compared to incongruent stimuli ( Koch and Allport, 2006 ). Several authors (e.g., Rosenbaum et al., 1986 ) have suggested a key role of motor programming in accounting for congruency effects. Accordingly, performance benefits in congruent compared to incongruent trials are due to the maintenance of the appropriate motor programming parameter in memory from trial to trial (e.g., if the same finger is used to respond), while motor parameters must be re-programmed in incongruent trials thereby increasing reaction times. Please note, that we take a broader perspective by considering parameters specifying motor programs as part of the task. In order to explain the congruency effect, it has been argued that incongruent stimuli activate both the response according to the currently relevant task rules (i.e., the relevant task set) and the response according to the currently irrelevant task rules (i.e., the irrelevant task set) of the competing task ( Kiesel et al., 2010 ). The congruency effect reflects the inability to shield the currently relevant task set from the currently irrelevant task set (see e.g., Dreisbach and Haider, 2009 ; Dreisbach and Wenke, 2011 ). Thus, while efficient task-set shielding should keep both task-sets distinct and prevent interference when alternating between tasks (i.e., decrease the congruency effect), less efficient task-set shielding would cause parallel processing and thus increase competition between tasks and responses arising in incongruent trails. To our knowledge, no study examined task-set shielding, as
processing to optimize taskswitching in terms of reduced switch costs and higher net multitasking efficiency compared to a strictly serial processing manner. A similar effect emerged at least on a descriptive level when comparing the dual-task efficiency of switchers and alternaters in the second experiment. However, neither in the first nor the second experiment the effects of overlapping processing were so strong to actually turn dual-task cost effects in dual-task benefits as would have been reflected in positive ODTPE scores. This was probably due to the fact that in both experiments overlapping processing was only applied in a limited number of trials only and hence could not fully compensate other cost effects. This certainly has prevented a more positive effect of this strategy on overall dual- task efficiency. One reason for this could be a lack of enough practice as needed to fully establish this strategy. Another reason might be the high similarity of the two tasks with respect to the specific processing resources (Wickens, 2002). This might have made task interleaving and overlapping processing particularly difficult and prone to negative side effects. However, it is remarkable that the subgroup of switchers in the second experiment nevertheless were able to realize actual time benefits associated with task switches in up to 47% of switch trials. This shows the potential benefits of a strategy combining task
TEMPORAL DISTINCTIVENESS IN EPISODIC TASK RETRIEVAL 47 The study at hand argues that the concept of temporal distinctiveness is important in the present context, so that this concept is transferred from serial memory to explaining RCI effects in taskswitching. In the context of cued taskswitching, it is assumed that the task-repetition benefit is largely driven by cue- triggered episodic retrieval of a task set, as the cue needs to retrieve an already activated task set (see also Mayr & Kliegl, 2003). Critically, it is assumed that in task repetitions this retrieval process benefits not only from the fact that the task set is already activated but also from an episodic retrieval component that refers to the previous encounter of the same task set. Hence, task-set retrieval is additionally modulated by the degree to which previous episodes referring to the same task set are retrievable. Successful episodic retrieval is influenced by different factors (e.g., Rugg & Wilding, 2000). The present work suggests that temporal distinctiveness is one further, albeit not the only, important variable that affects the degree to which episodic memory traces of previous task episodes are retrievable. By applying this idea to the effect of RCI manipulations in taskswitching, it is assumed that temporal distinctiveness can be defined as the ratio of the current RCI to the previous RCI 1 and that this ratio largely determines the RCI effect. Thus, RCI effects influence temporal distinctiveness, which in turn affects episodic retrieval. This idea of temporal
The taskswitching paradigm applied during behavioral training was similar to the one of Karbach and Kray (2009) , albeit a cued paradigm variant. Different stimulus material and different tasks were used in each session given that training variability may enhance the scope of training transfer ( Karbach and Kray, 2009 ). The order of the different training-task sets across sessions was kept constant for all participants. Participants were asked to categorize pictures 1 by means of a left or right response button according to a semantic task (e.g., fruit or vegetable?), or a perceptual task (e.g., small or large size?). Critically, response formats were overlapping for both task sets to increase the demand on executive control (cf. Kray and Fehér, 2017 ). The semantic and perceptual task sets were (see also Figure 1): in the first session, transportation (car or plane?) and number task (single or double?); in the second session, hobby (music or sports?) and color task (blue or orange?); in the third session, animal (fish or bird) and direction task (right or left?) in the fourth session, plant (leaf or flower?) and chromaticity task (colored or black and white?); in the fifth session, clothing (hat or shoe?) and texture task (dotted or squared?); in the sixth session, landscape (building or tree?) and orientation task (upright or rotated?); in the seventh session, gadget (toy or 1 Stimuli were mainly retrieved from the databases of Snodgrass and Vanderwart
Brüning, J., & Manzey, D. (2017). Flexibility of individual multitasking strategies in task-switching with preview: are preferences for serial versus overlapping task processing dependent on between-task conflict? Psychological Research, 82(1), 92–108. https://doi.org/10.1007/s00426-017-0924-0 This is a post-peer-review, pre-copyedit version of an article published in Psychological Research. The final authenticated version is available online at: http://dx.doi.org/10.1007/s00426-017-0924-0.
Hence, small gender differences in multitasking abilities across women and men in the used task-switching and dual-task paradigms cannot be excluded based on the present study. Moreover, the present study does not allow any conclusions about gender differences in other multitasking situations, which for example call for more planned and future-oriented strate- gies or involve offloading of spatial abilities [ 19 ]. However, considering the good power of the present study to detect even medium-to-large gender differences, the present findings strongly suggest that there are no substantial gender differences in multitasking performance across task-switching and dual-task paradigms, which predominantly measure cognitive control mechanisms such as working memory updating, the engagement and disengagement of task sets, and inhibition.
from developmental disturbances or even disorders associated with executive control. For instance, accumulating evidence has documented that impairments in executive-control processes lie at the core of behavioral symptoms in children diagnosed with attention-deficit/hyperactivity disorder (Crosbie et al., 2013; Kofler, Rapport, Bolden, Sarver, & Raiker, 2010; Martinussen, Hayden, Hogg- Johnson, & Tannock, 2005; Rapport et al., 2009; Tillman, Eninger, Forssman, & Bohlin, 2011; and Willcutt, Doyle, Nigg, Faraone, & Pennington, 2005). ADHD is a heritable neurodevelopmental disorder with reported rates of prevalence varying between 2.2 and 17.8 % worldwide in school-aged children and adolescence (for a review, see Skounti, Philalithis, & Galanakis, 2007) and associated with chronicity over the lifespan if not treated adequately. The core behavioral symptoms in ADHD are characterized by poor regulation of attention, impulsivity, and physical activity. 13 Such impairment of behavior dramatically increases the risk of negative long-term outcomes across various domains, such as addictive or antisocial behavior, academic underachievement, occupational difficulties, or obesity (for a review, see Young, Fitzgerald, & Postma, 2013). The key driving mechanism of executive control impairments underlying ADHD symptoms in childhood is debated to be a major deficit in inhibition (e.g., Barkley, 1997; Quay, 1988) or in working memory (Kofler, Rapport, Bolden, & Altro, 2008; Martinussen et al., 2005; Rapport et al., 2008), that both are exacerbated in task- switching situations (Cepeda et al., 2000). The latter finding has been attributed to specific performance impairments in ADHD children when several executive components are required to act together in an interactive rather than an additive fashion, such as in an un-cued, alternating-runs variant of taskswitching (i.e., high memory load) coupled with ambiguous stimuli and response formats (i.e., high inhibition load; e.g., Wu, Anderson, & Castiello, 2006).
Modality compatibility denotes the match between sensory stimulus modality and the sensory modality of the anticipated response effect (for example, vocal responses usually lead to auditory effects, so that auditory–vocal stimulus–response mappings are modality-compatible, whereas visual–vocal mappings are modality incompatible). In taskswitching studies, it has been found that switching between two modality-incompatible mappings (auditory-manual and visual–vocal) resulted in higher switch costs than switching between two modality-compatible mappings (auditory–vocal and visual-manual). This finding suggests that with modality-incompatible mappings, the anticipation of the effect of each response primes the stimulus modality linked to the competing task, creating task confusion. In Experiment 1, we examined whether modality- compatibility effects in taskswitching are increased by strengthening the auditory–vocal coupling using spatial-verbal stimuli relative to spatial-location stimuli. In Experiment 2, we aimed at achieving the same goal by requiring temporal stimulus discrimination relative to spatial stimulus localisation. Results suggest that both spatial-verbal stimuli and temporal discrimi- nation can increase modality-specific task interference through a variation of the strength of anticipation in the response-effect coupling. This provides further support for modality specificity of cognitive control processes in taskswitching.
Despite the commonalities between task-switching and dual-task- ing, there are only a few studies and theoretical considerations about how the performance costs in task-switching and dual-tasking might be related (e.g., Band, Jolicoeur, Akyürek, & Memelink, 2006 ; Oriet & Jolicoeur, 2003 ; Pashler, 2000 ; Sigman & Dehaene, 2006 ). For instance, by implementing T1-T2 switches and T1-T2 repetitions into dual-task trials, Band and van Nes ( 2006 ) as well as Lien et al. ( 2003 ) observed, in addition to a PRP effect, switch costs in T2. The cost was comparable across SOAs. Lien et al. ( 2003 ) concluded from the switch costs in T2 that time is needed for processes involved in the disengagement of T1 and the engagement of T2, and thus, that the shifting component is in- volved in dual-tasking (see also Liepelt, Strobach, Frensch, & Schubert, 2011 , for similar conclusions derived from studies on practice effects in dual-tasking; for a review, see, e.g., Strobach & Schubert, 2017 ). They argued that the missing absorption of switch costs in T2 during the waiting period for the availability of the response-selection bottleneck (i.e., additivity of SOA and T1-T2 transition) suggests that these proc- esses, at least partially, occur after response selection for T1 and before response selection for T2, and modified the response-selection bot- tleneck accordingly.
Despite the commonalities between task-switching and dual-task- ing, there are only a few studies and theoretical considerations about how the performance costs in task-switching and dual-tasking might be related (e.g., Band, Jolicoeur, Akyürek, & Memelink, 2006 ; Oriet & Jolicoeur, 2003 ; Pashler, 2000 ; Sigman & Dehaene, 2006 ). For instance, by implementing T1-T2 switches and T1-T2 repetitions into dual-task trials, Band and van Nes ( 2006 ) as well as Lien et al. ( 2003 ) observed, in addition to a PRP eﬀect, switch costs in T2. The cost was comparable across SOAs. Lien et al. ( 2003 ) concluded from the switch costs in T2 that time is needed for processes involved in the disengagement of T1 and the engagement of T2, and thus, that the shifting component is in- volved in dual-tasking (see also Liepelt, Strobach, Frensch, & Schubert, 2011 , for similar conclusions derived from studies on practice eﬀects in dual-tasking; for a review, see, e.g., Strobach & Schubert, 2017 ). They argued that the missing absorption of switch costs in T2 during the waiting period for the availability of the response-selection bottleneck (i.e., additivity of SOA and T1-T2 transition) suggests that these proc- esses, at least partially, occur after response selection for T1 and before response selection for T2, and modiﬁed the response-selection bot- tleneck accordingly.
3.1.1 The structural boundaries of modality compatibility
Previous research on modality-compatibility effects in taskswitching suggests indisputably that modality compatibility influences switch costs. Moreover, the observed modality-compatibility effects could be attributed to the central component of cognitive processes, due to the implemented design in which stimulus and response modalities were equated across conditions. Note that this means that participants switched between two modality-incompatible tasks in one condition and between two modality-compatible tasks in the other condition (e.g., Stephan & Koch, 2010, 2011, 2016). According to the ideomotor theory, switching between modality incompatible tasks evoke between-task crosstalk during the response selection process, but reduced between-task crosstalk could be observed with modality-compatible tasks. Also, ideomotor theory indicates that between-task crosstalk is supposed to be reduced if at least one task from the currently relevant task sets is modality compatible. This question was addressed in Study I for which we developed an experimental design with mixed modality mappings in order to investigate all possible combinations of modality mappings. The only way to do this was by keeping either the stimulus modality constant while switching between response modalities or by holding the response modality constant while switching between stimulus modalities. Note that there have been a number of studies that investigated the influence of either the response-modality switches (e.g., Philipp & Koch, 2005, 2010, 2011; Philipp, Weidner, Koch, & Fink, 2013) or the stimulus-modality switches (Hunt & Kingstone, 2004; Kreutzfeldt, Stephan, Sturm, Wilmes, & Koch, 2015; Lukas, Philipp, & Koch, 2010a; Murray, de Santis, Thut, & Wyle, 2009; Sandhu & Dyson, 2012) but these studies did not focus on the role of modality compatibility.
Research on remote tower control solutions for small airports raised the question whether it is possible for an air traffic controller to control multiple airports simultaneously. This simultaneous control would require the air traffic controller to switch between task sets of two airports. Therefore it is important to analyse the factors influencing taskswitching in these dynamic multiple task environments and how they are affecting the air traffic controller.
The important role of predictability for task-switching was amongst others shown by Monsell and Mizon (2006). With a task-cueing paradigm and a 2:1 cue-target mapping, switch- specific preparation effects can be obtained when the probability for a task-switch is reduced to 25% and by that the predictability for repetition trials is increased automatically. However, one always has to be cautious when interpreting difference scores in terms of task- or attention switch costs, as a reduction of these might either be produced by a RT reduction in switch-trials or by strong priming in repetition trials, which does not benefit from increased preparation time. When the probability for a task-switch is only 25%, participants have a high benefit of anticipating task- repetitions for upcoming trials and therefore RT in repetition trials gets shorter and does not benefit greatly from increased preparation time. Instead, a rare task-switch “disturbs” participants relatively more and therefore benefits from longer time available for preparation. When, in contrast, the switch-probability is as high as 75%, switch-specific preparation might take place, but is not revealed by the data because the rare task-repetition trials then might be slower due to their low predictability and produce longer RT than in trials with a short preparation interval. The alternating-runs paradigm instead is void of this switch-probability confound, and nevertheless provides an entirely predictable switch-sequence. Hence, a switch-specific reduction of RT in the alternating-runs paradigm, as found for auditory attention-switching in our third experiment may show true switch-specific preparation.
Switching experiments in memory cells are a powerful tool to investigate the ki- netics and electronic effects during the phase transitions on short timescales. They can be complemented with high sensitive electric measurements on long timescales. Temperature dependent resistance measurements on unstructured thin films pro- vide insight in both crystallization temperatures, thermal activated carrier trans- port, and allow investigation of the drift mechanism in amorphous phase-change materials (chapter 6). These properties could be investigated using memory cells, but the highly demanding production process would hinder a fast material screen- ing, and the two-point measurement technique is not optimized for recording small changes in resistance. Therefore, amorphous as deposited phase-change layers on insulating substrates and the van-der-Pauw method for sheet resistance measure- ments [Pauw:1958] were chosen for this task.
A qualitatively similar picture is found in the case of a rotating field with angular frequency directed opposite to the vortex polarity: wp < 0. There exist different regimes in the vortex dynamics . In the range of frequencies close to the frequency of the orbital vortex motion ( W ~1 GHz the vortex demonstrates a finite motion in a ) region near the disk center along a quite complicated tra- jectory and does not switch its polarity. These cycloidal vortex oscillations are similar to those in Ref. , and correspond to the excitation of higher magnon modes dur- ing the motion. If w >> W, then for weak fields the vortex motion can be considered as a sum of two constituents: (i) the gyroscopic orbital motion as without field, and (ii) cy- cloidal oscillations caused by the field influence. For the case wp < 0 the direction of the cycloidal oscillations coin- cides with the direction of the field rotation, while the di- rection of the gyroscopic orbital motion is opposite to it. For stronger fields the vortex motion becomes more com- plicated and its average motion can be directed even op- posite to the gyroscopical motion (this situation is shown in Fig. 1). The irreversible switching of the vortex polar- ity can be excited in a specific range of parameters ( , ) w b , with typical frequencies about 10 GHz and intensities about 20 mT . The mechanism of the switching is discussed below.
There are some different proposed models/theories trying to explain the resistive switching in MIM devices based on electrical effects. Electronic charge injection and/or charge displacement is seen as one origin of the switching. As early as 1967, Simmons and Verderber proposed a charge–trap model which can explain the different resistance states of a MIM cell by the modification of the electrostatic barriers due to trapped charges . According to this model, charges are injected by Fowler–Nordheim tunneling at high electric fields and subsequently trapped at sites such as defects or metal nanoparticles in the insulator. This model was later modified in order to incorporate charge trapping at interface states, which is thought to affect the adjacent Schottky barrier at various metal/semiconducting perovskite interfaces . Another proposed explanation for perovskite–type oxides is based on the insulator–metal transition (IMT) . In this model, the electronic charge injection acts like doping, which induces an IMT [39, 40].