Information

What is the study with the dual-task experiment that involved a colour-wheel change detection task?

What is the study with the dual-task experiment that involved a colour-wheel change detection task?


We are searching data for your request:

Forums and discussions:
Manuals and reference books:
Data from registers:
Wait the end of the search in all databases.
Upon completion, a link will appear to access the found materials.

This may be a long-shot, but I'm looking for a paper that I vaguely remember reading a few years ago. Unfortunately I can't remember many details about the paper or its content, and thus my multiple attempts at keyword searches have failed. I'm posting this question in the hope that someone recognises the few details I do remember.

The paper presented more than one experiment, but the one I'm trying to recall involved a colour wheel/pie displayed in the centre of the screen. The slices of the pie were each a different colour and the colours would change a few times throughout each trial. There were only a small number of slices/colours. The observer's task was to count the number of times the colours changed. While doing this the observer also had to perform another task, for which the stimuli were shown on the rest of the display, in peripheral vision. I can't remember what the other task was, though it may have been a multiple object tracking task or a visual search task. The paper may have been about the effect of attentional load on working memory, though I'm not sure about that.


Your description sounds a little bit like Experiment 3 from

Hollingworth, A., Richard, A. M., Luck, S. J. (2008) Understanding the function of visual short-term memory: Transsaccadic memory, object correspondence, and gaze correction. Journal of Experimental Psychology: General, Vol 137(1), 163-181. PDF

See Fig. 7 on page 172 for a picture of the display utilized in this experiment.


I eventually tracked it down by going through every single result of every possible relevant search I could think of.

Makovski, T., & Jiang, Y. V. (2009). The Role of Visual Working Memory in Attentive Tracking of Unique Objects. Journal of Experimental Psychology: Human Perception and Performance, 35(6), 1687-1697.

It's available for free on the NIH site. It was experiment 4. Figure 5 gives an illustration of the stimuli.


2. Results

The central findings were that the sentence listening task reliably degraded driving performance, and in addition, it resulted in decreases in activation in key regions that underpin the driving task, as further quantified below.

2.1 Behavioral Measures

Participants performed the sentence comprehension task at a 92% accuracy level (SD = 0.06 %), confirming that they were attending to the auditory stimuli in the driving with listening condition. The behavioral measures indicated reliably more road maintenance errors and larger root mean squared (RMS) deviation from an ideal path in the driving with listening condition. Mean road maintenance errors (hitting the berm) increased from 8.7 (SD = 9.7) in the driving alone condition to 12.8 (SD = 11.6) in the driving while listening condition (t(28) = 2.22, p < .05). The mean RMS deviation from the ideal path increased from 2.48 to (SD = 0.51) in the driving alone condition to 2.64 (SD = 0.56) in the driving while listening condition (t(28) = 2.79, p < .01). Both of the measures of driving accuracy are essentially continuous visuo-spatial tracking measures rather than reaction time measures of hazard avoidance. A meta-analysis (Horrey and Wickens, 2006) of 16 behavioral studies of dual-task driving concluded that the costs associated with cell phone conversations are even larger for reaction time tasks than for tracking tasks, so our study may be underestimating the behavioral impact of a secondary task on driving.

2.2 Functional Imaging Measures

Group-level random-effects analysis indicated that the driving task when performed alone produced large areas of activation (compared to fixation) in bilateral parietal and occipital cortex, motor cortex, and the cerebellum, as shown in Figure 2A . Three clusters of activation survived correction for multiple comparisons (p < .05). The largest cluster (39,504 voxels) had its peak activation in the left supplementary motor area (t(28) = 12.00, at Montreal Neurological Institute (MNI) coordinates -6 -18 64), but the activation extended to left and right primary motor areas, the left and right parietal lobe, the left and right occipital lobe, and into bilateral regions of the cerebellum. A second cluster (1,791 voxels) had a peak in the left thalamus (t(28) = 8.72 at MNI coordinates -14 -22 2) but extended into other left subcortical structures including the putamen, pallidum, caudate, and hippocampus, and also left cortical areas of the insula, inferior frontal gyrus, and middle frontal gyrus. The final cluster (429 voxels) had its peak in the the right hippocampus (t(28) = 7.71 at MNI coordinates 22 -30 -8) and extended into the right thalamus, and right cortical areas of the parahippocampal and lingual gyri.

Whole-brain voxel-wise random effects statistical parameter maps of each condition contrasted with the fixation baseline thresholded at p < .0001 with an 81-voxel extent threshold (resulting in a cluster-level threshold of p < .05 after correction for multiple comparisons). Similar areas of activation are present in both conditions but with additional language-related activity in temporal and inferior frontal areas (yellow ovals).

When sentence listening was combined with the driving task, the same network of driving-related areas were activated, as shown in Figure 2B . For the contrast between driving with listening and the fixation baseline, the largest cluster of activation (47,911 voxels) had a peak in the right middle occipital gyrus (t(28 = 12.43 at MNI coordinates 28 -96 4) but extended to the same areas found in the contrast of driving alone with fixation left and right supplementary and primary motor areas, left and right parietal lobes, left and right occipital lobes, and bilateral areas of the cerebellum. As expected, the addition of the listening task gave rise to activation in additional areas that underpin the sentence processing task, namely bilateral temporal and left inferior frontal regions. The largest cluster of activation extended into the left inferior frontal gyrus, and also into the left temporal language area (see the left panel of Figure 2B ). In addition, a cluster of 3,022 voxels was reliably active in the homologous region of the right temporal lobe (peak t(28) = 10.99 at MNI coordinates 50 -24 -6). A final small cluster of activation (185 voxels) was found in the right frontal lobe with a peak in the middle frontal gyrus (t(28) = 6.14 at MNI coordinates 24 52 6).

If processing spoken language draws attentional/brain resources away from the task of driving, one would expect a decrease in activation in the brain areas that underpin the driving task. The findings clearly supported this prediction. Informal comparison of Figure 2A and Figure 2B suggests that the driving-related activation in bilateral parietal cortex decreased with the addition of the sentence listening task. Direct random-effects statistical comparison of the driving alone condition with the driving with listening condition confirms this suggestion (see Figure 3 and Table 1 ). A number of bilateral occipital and parietal areas showed greater activation in the driving alone condition relative to the same condition performed with the sentence listening task, as shown in Figure 3A and in Table 1 . As expected, driving with listening resulted in more activation than driving alone in bilateral temporal language areas and the left inferior frontal gyrus, as shown in Figure 3B and in Table 2 . There was also greater activation in the right supplementary motor area in this contrast, possibly due to the addition of the requirement to respond to the sentence comprehension task with the left hand.

Whole-brain voxel-wise random effects statistical parameter maps of direct contrasts between the two conditions thresholded at p < .0001 with a 81-voxel extent threshold (resulting in a cluster level threshold of p < .05 after correction for multiple comparisons). The top panel indicates that parietal and superior extrastriate activation decreases with the addition of a sentence listening task (blue circle). The bottom panel shows that the addition of a sentence listening task results in activation in temporal and prefrontal language areas (yellow ovals).

Table 1

Areas of greater activation for Driving Alone than Driving with Listening

MNI Coordinates
Location of Peak activationCluster Sizet(28)xyz
L Supramarginal Gyrus1667.13-56-3636
R Superior Parietal Lobe2,0206.810-8252
L Superior Parietal Lobe1395.8-28-5458
L Inferior Parietal Lobe1545.55-34-4238
L Superior Occipital Gyrus1825.49-26-8826

Note: Cluster size is in 2 × 2 × 2 mm voxels. L = Left, R = Right.

Table 2

Areas of greater activation for Driving with Listening than Driving Alone

MNI Coordinates
Location of Peak activationCluster Sizet(28)xyz
L Middle Temporal Gyrus4,55210.87-56-12-6
Right Superior Temporal Gyrus2,5239.8250-204
L Inferior Frontal Gyrus4979.33-442026
R Supplementary Motor1,0557.0022462

Note: Cluster size is in 2 × 2 × 2 mm voxels. L = Left, R = Right.

Anatomical regions of interest (ROIs) defined a priori were used to directly compare the activation levels (percentage change in signal intensity relative to fixation) in the two conditions. There were large, reliable decreases in areas involved in the spatial processing associated with driving. The decrease from single to dual task was 37% for the spatial areas (F(1, 28) = 29.38, p < .0001. Table 3 shows the mean percentage change in signal intensity for each of the anatomically defined regions of interest examined in the driving alone and driving with listening conditions. Most of the parietal areas associated with spatial processing individually showed a reliable decrease in activation when the sentence comprehension task was added, with the largest decreases found in the right parietal lobe. Table 3 also groups the anatomical areas based on function, and Figure 4 aggregates the results for each of these groupings. As shown in Figure 4 , the spatial areas show a large decline in activation in driving with listening compared to driving alone the visual, motor, and executive areas show no reliable decrease and the language areas show a large increase.

The percentage change in signal intensity for five functional groupings (networks) of cortical areas. The component regions of each network are those specified in Table 3 . The driving-related activation in spatial processing areas significantly decreases with the addition of the sentence listening task. The addition of the sentence listening task significantly increases language area activation. Error bars show the standard error of the mean.

Table 3

Mean percentage change in signal intensity in anatomical regions of interest (ROI)

Region of InterestDriving AloneDriving with ListeningF(1, 28)
Spatial Areas
L Intraparietal Sulcus0.315 >0.2318.14 *
R Intraparietal Sulcus0.400 >0.26714.28 **
L Inferior Parietal Lobe0.461 >0.3485.67 *
R Inferior Parietal Lobe0.0830.0113.64
L Superior Parietal Lobe0.239 >0.15810.23 *
R Superior Parietal Lobe0.226 >0.12014.01 **
L Superior Extrastriate0.337 >0.2346.63 *
R Superior Extrastriate0.374 >0.2469.25 *
All Spatial Areas0.258 >0.16329.38 **
Visual Sensory/Perceptual Areas
Calcarine Sulcus0.1890.1431.56
L Inferior Extrastriate0.2670.2161.52
R Inferior Extrastriate0.3060.2442.66
L Inferior Temporal Lobe0.1380.1080.17
R Inferior Temporal Lobe0.1790.1091.20
L Inferior Temporal Lobe0.1110.1400.05
R Inferior Temporal Lobe0.1490.1290.02
All Visual Areas0.1910.1561.39
Motor/Pre-Motor Areas
Supplementary Motor Area0.2120.2441.73
L Precentral Gyrus0.4290.3801.68
R Precentral Gyrus0.2220.1960.76
All Motor Areas0.2880.2730.32
Executive Function Areas
L Middle frontal Gyrus0.1080.0920.23
R Middle Frontal Gyrus0.1130.0761.34
Anterior Cingulate-0.085-0.0960.18
Superior Medial Frontal-0.085-0.0960.18
All Executive Areas0.0350.0300.07
Language Areas
L Ant. Superior Temporal0.043 <0.39942.45 **
R Ant. Superior Temporal0.076 <0.39121.95 **
L Pos. Superior Temporal-0.024 <0.21437.98 **
R Pos. Superior Temporal-0.012 <0.0774.29 *
L Pars Triangularis0.114 <0.25612.64 **
R Pars Triangularis0.081 <0.1616.01 *
L Pars Opercularis0.1360.1781.36
R Pars Opercularis0.1800.1670.18
L Insula0.0740.0900.21
R Insula0.0360.0270.07
All Language Areas0.070 <0.19664.43 **

Note: inequality signs indicate the direction of a statistically reliable difference between Driving Alone and Driving with Sentence Listening. L = Left, R = Right.

Although the visual areas show a trend toward a decrease in activation between the driving alone condition and the driving with listening condition, this decrease was not reliable for any of the areas considered individually or for the aggregate measure of visual activation. However, more superior areas of the right and left occipital lobe did show significantly less activation for the driving with listening condition in the voxel-wise whole brain contrasts (see Figure 3A ). These areas have been grouped with the spatial processing areas in Table 3 and Figure 4 , due to their proximity to the parietal lobes and their role in the dorsal visual stream, but this grouping is perhaps somewhat arbitrary. The data indicate that while primary visual areas show no effect of the multitasking in this study, some secondary visual areas do decrease their activation.

In frontal areas associated with executive function, including dorsolateral prefrontal cortex and anterior cingulate, one might expect that the need to coordinate the processing in the two tasks would lead to increased activation, as D'Esposito et al. (1995) reported. However, note the previous distinction between performing two tasks concurrently (such as driving and sentence listening) versus rapidly switching between two tasks (such as the dual tasks studied by D'Esposito et al.). Unlike the findings of increased activation in prefrontal areas for task switching, these prefrontal regions showed an equivalent percentage change in signal intensity for the driving alone and driving concurrently with sentence listening conditions. This finding indicates that not all multitasking requires additional executive functioning.

As expected, there was an overall increase in the percentage change in signal intensity in language areas when the comprehension task was added to the driving task. This increase was prominent in bilateral primary and secondary auditory areas of the temporal lobe and in the pars triangularis region of Broca's area in the left hemisphere and the homologous region of the right hemisphere, as indicated in Table 3 . There was a slight trend toward a greater percentage change in signal in left pars opercularis, consistent with the results of the voxel-wise analysis, but not in right pars opercularis.

The finding of decreased parietal activation for the driving with listening condition was also found when the volume of activation rather than the percentage change in signal intensity was considered. For this analysis, the number of voxels reliably activated in the a priori spatial anatomical ROIs was computed for each participant at t > 4.90 (corresponding to a within-participant height threshold of p < 0.05, corrected for multiple comparisons) for the contrast of each condition with the fixation baseline. In the spatial areas, as identified in Table 3 , the mean total number of activated voxels decreased from 1,653 (SE = 103) to 1,195 (SE = 103) from the driving alone condition to the driving with listening condition, (F(1, 28) = 41.65, p < .0001).


Executive function and motor control deficits adversely affect gait performance with age, but the neural correlates underlying this interaction during stair climbing remains unclear. Twenty older adults (72.7 ± 6.9 years) completed single tasks: standing and responding to a response time task (SC), ascending or descending stairs (SMup, SMdown) and a dual-task: responding while ascending or descending stairs (DTup, DTdown). Prefrontal hemodynamic response changes (∆HbO2, ∆HbR) were examined using functional near-infrared spectroscopy (fNIRS), gait speed was measured using in-shoe smart insoles, and vocal response time and accuracy were recorded. Findings revealed increased ∆HbO2 (p = 0.020) and slower response times (p < 0.001) during dual- versus single tasks. ∆HbR (p = 0.549), accuracy (p = 0.135) and gait speed (p = 0.475) were not significantly different between tasks or stair climbing conditions. ∆HbO2 and response time findings suggest that executive processes are less efficient during dual-tasks. These findings, in addition to gait speed and accuracy maintenance, may provide insights into the neural changes that precede performance declines. To capture the subtle differences between stair ascent and descent and extend our understanding of the neural correlates of stair climbing in older adults, future studies should examine more difficult cognitive tasks.

Stairs have been identified as a common source of injurious falls amongst older adults [1]. From a neural perspective, stair ambulation may be increasingly challenging with advancing age due to structural and metabolic changes in the prefrontal cortex (PFC) [2,3]. As such, cognitive functions, such as executive functions, which are supervised by the PFC, may be less efficient at processing challenging cognitive demands. Studies that have examined the relationship between executive functions and gait in older adults have demonstrated that executive function deficits are associated with declines in gait performance [4,5,6,7]. Compared to overground walking, stair ambulation may be highly demanding because it involves elements of dynamic balance, lower body strength and attention [8,9]. Additionally, for both ascent and descent, motor planning begins prior to reaching the first step and continues thereafter to ensure precise foot placement and the proper integration of sensory and visual information [10,11,12]. Interestingly, older adults who report a fear of falling may increase their handrail usage to ensure stability on stairs [9,13]. Older adults equally suggest that stair ascent and descent pose separate challenges [12,13]. For example, greater balance compensation is required during stair descent, whereas ascent involves greater physical exertion to counteract gravitational forces [1,10]. These differences may also account for an increased incidence of falls during stair descent and slower gait speeds during stair ascent [12].

The dual-task paradigm may be used to better understand executive functioning during stair ambulation by examining an individual’s capacity to manage two tasks simultaneously [14,15]. When two overlapping processes compete for the same cognitive resources, performance on one or both tasks may suffer [16]. This is known as a dual-task cost which can be assessed using a variety of cognitive and motor performance measures. For example, older adults have demonstrated reduced vocal response times during a reaction time task while ambulating stairs [17]. Similar findings were obtained during a standing balance task, whereby decreasing the base of support and increasing instability led to slower response times in older adults [18]. Other studies have focused primarily on stair descent, where older adults decreased their gait speed compared to overground walking [19] and while responding to a mental arithmetic task [20]. Accuracy scores proved to be more dependent on cognitive task difficulty, such that higher scores were achieved during response time tasks [21] compared to serial subtractions [22] and working memory tasks [23] during overground walking. Thus, performance costs during dual-tasks may offer an indirect measure of executive functioning and allow for further insight into the neural mechanisms involved in stair ambulation.

In contrast, our knowledge of direct measures of neural activation on stairs has been limited by the restrictive nature of most neuroimaging techniques. However, functional near infra-red spectroscopy (fNIRS) has emerged as an important tool to measure neural activation because it is robust to motion artifacts and does not limit participant mobility [24,25]. fNIRS uses the principle of neurovascular coupling to measure the changes in cerebral oxygenation (∆HbO2) and deoxygenation (∆HbR) following a neural stimulus. When a series of vascular events are initiated to mitigate the increased metabolic demand of oxygen, the changes in cerebral blood flow and oxygen metabolism can be coupled and used as a neurophysiological marker to detect changes in brain activation [26]. PFC activity during overground walking and obstacle negotiation has been well documented in the literature, such that greater motor task complexity is associated with greater PFC activation [24,27,28]. Similarly, studies have demonstrated that older adults exhibit greater PFC activation during dual-task walking compared to walking alone [15,29,30]. This is in line with the revised scaffolding theory of aging and cognition (STAC-r) which suggests that older adults may adapt to age-related neurodegeneration by recruiting additional neural networks [31]. Therefore, stronger and bilateral recruitment of the PFC can be expected in older versus younger adults to maintain a high level of cognitive and motor performance during stair ambulation.

This study builds upon previous behavioral work that examined changes in gait speed and response time performance in older adults during stair negotiation [17,20]. The purpose of this study was to evaluate the hemodynamic response (∆HbO2 and ∆HbR) and performance (gait speed, response time and response accuracy) changes in older adults under single and dual-task stair ascent and descent. It was hypothesized that dual-tasks are more complex than single tasks and will invoke worse cognitive (e.g., response time and response accuracy) and motor (e.g., gait speed) performance due to a greater reliance on executive functions and the PFC. Similarly, stair ascent is more physically demanding, which may result in worse motor performance, whereas stair descent requires more planning and conscious attention, which may lead to worse cognitive performance. From a neural perspective, greater ∆HbO2 and ∆HbR are expected during the dual- compared to single tasks in line with the STAC-r neural compensation theory [31]. In addition, stair descent may require greater cognitive control than stair ascent, which is expected to elicit a greater hemodynamic response change in the PFC.


Materials and Methods

Participants

Twenty healthy, right-handed adults (11 females), aged 20�, participated in the study. All participants had normal or corrected to normal vision. Additionally, all participants were not expert video game players, as defined by having less than 2 h of video-game usage per month in the past 2 years. All participants gave informed consent and were compensated for their participation.

Stimuli and Procedure

The dual-task paradigm consisted of a lane change driving task and an image discrimination task. The driving environment was designed in the Unity 3D game engine. Participants sat at a distance of 50 cm from a 22″ LG monitor with a refresh rate of 60 Hz and a resolution of 1,920 × 1,080 and responded to the tasks using a computer keyboard.

The driving environment consisted of a three-lane, desert road, without left/right turns or inclining/declining hills. Driving stimuli, composed of two rows of traffic cones (three cones in each row Figure 2A ), were presented on the two sides of one of the lanes in each trial, and the participants had to immediately redirect the car to the lane with the cones and pass through the cones. The space between the two rows of cones was such that the car could easily pass through them without collision. The cones were always presented in the lanes immediately to the left or immediately to the right of the car’s lane so that the participants had to change only one lane per trial. The lane change was done gradually: the participant had to hold the corresponding key to direct the car in between the two rows of cones, and then release the key when the car was situated correctly. Any early or late key press or release would cause a collision with the cones and a performance loss in that trial. The fixation cross was jittered for 100 ms to provide online feedback in case of a collision with the traffic cones. The participants were instructed not to change lane before the cones appeared. Trials in which participants changed lane before the presentation of the cones were considered false and removed from the analysis. Using this method, we could divide a continuous driving task into individual trials with predetermined onset and ends. At the beginning of the block, participants speeded up to 80 km/h using the “up” arrow key with the middle finger of the right hand. During the block, the speed was kept constant, and the lane change was performed by pressing the right and left arrow with the middle and index fingers of their right hand, respectively. For the image discrimination task, a single image of either a scene or a face was presented for 150 milliseconds centered at 2° eccentricity above the fixation cross ( Figure 2B ). The size of the image was 2.5° of visual angle. Participants pressed the “x” and “z” keys on the computer keyboard with the middle and index fingers of their left hand to determine whether the image was a face or a scene, respectively. The images were pseudo-randomly selected from a set of 864 images of scenes and 435 images of faces. We selected only natural scenes and neutral faces. If participants responded incorrectly, the green fixation cross turned red, and if they responded late, it turned orange for 100 ms. The length of each trial was 3 s, and the inter-trial interval varied randomly from 0.5 to 1.5 s. For the first trial in each block, the onset of the trial was set to 2 s after the beginning of the block. The end of the trial was set to when the rear end of the car reached the end of the set of traffic cones.

Dual-task paradigm. (A) A sample display showing the driving stimulus consisting of two rows of traffic cones in the middle driving lane. The cones were randomly presented in each lane, and participants had to drive through them without collision. (B) A sample display showing an image discrimination presented above the fixation point. Participants determined if the image was a face or a scene. (C) The sequence of events for a sample trial in which the image task was presented first (left), and another in which the driving task was presented first (right). The inter-trial interval (ITI) varied between 0.5 and 1.5 s. The image lasted for 150 ms, and the cones were presented 30, 100, 300, or 600 before or after the image. Participants had to perform a lane change immediately after the appearance of the cones, and an image discrimination task immediately after the presentation of the image.

The experiment consisted of two different conditions: (1) “Predictable” task order condition, and (2) “Unpredictable” task order condition. In two experimental conditions, the two tasks were presented with eight possible SOAs (�, �, �, �, +30, +100, +300 and +600 ms). In the negative SOAs, the image discrimination was presented first (image-first, Figure 2C ), and in the positive SOAs, the lane change was presented first (lane change-first, Figure 2C ). In the Predictable conditions, the order of the presentation was fixed, so that in two of the four blocks, the driving task was presented first, and in the other two, the image discrimination task was presented first. In the Unpredictable condition, the order of the presentation of the two tasks was not predictable in each trial. Trials with driving as the first task were interleaved with trials with the image discrimination as the first task. Before the start of each block, participants were informed about the type of the block.

In addition to the dual-task conditions, participants performed two single-task conditions: (1) single driving task and (2) single image discrimination task. In the single-task conditions, both the lane change and image stimuli were presented, but the participant only responded to one of them, ignoring the other. In the single image discrimination condition, the driving was on autopilot, and participants only responded to the images. In the single lane change condition, participants performed the lane change task and ignored the images.

Participants were told to focus on the fixation cross at the center of the page and respond to each task as fast as possible. At the end of each block, participants were informed about their performance on each task as well as their total performance. The performance in the driving task was calculated as the percentage of trials in which the participant passed through the cones without collision. The performance in the image discrimination task was calculated as the percentage of correct identifications.

Participants completed four blocks of 64 trials for each dual-task condition and two blocks of 32 trials for each single-task condition. There was a 1-min interval between blocks and a 5-min break after finishing all the blocks in each condition. The order of the blocks was counterbalanced across participants.

Before performing the main experiment, all participants performed a block of 20 trials for every single-task. If their accuracy was 80% or higher, they proceeded to the main experimental blocks. Otherwise, they repeated blocks of 40 trials for each task until they reached 80% accuracy. After the single-task training, participants performed the dual-task training block. The dual-task training was similar to the single-task training block, with the difference that if after 20 trials, the dual-task performance did not reach the 75% threshold, the training was repeated with blocks of 50 trials.

Drift Diffusion Model Fitting

To investigate if the two tasks were processed serially, or in parallel we used a DDM in which each trial was modeled as a combination of a non-decision time and a decision time consisting of a random drift towards decision bound ( Figure 1 ). Model parameters consisted of: (1) parameter z denoting the starting point of the decision process, (2) parameter a denoting the decision threshold, (3) parameter v representing the speed of information accumulation or drift rate, and (4) parameter t0 denoting the non-decision time pertaining to the combination of all other times in the trial excluding the drift-diffusion time. The DDM was implemented in the current study, by fitting the parameters z, a, v, and t0. We modified the DDM, so that z and a were independent of SOA, and v and t0 were dependent on SOA. Therefore, in the modified DDM, four values were fit for the parameter v and four values for the parameter t0 corresponding to the four SOAs, one value for the parameter a and one value for the parameter z across all SOAs.

We used the Fast-dm package, developed by Voss and Voss (2007), for model fitting. Fast-dm is a package for fast drift-diffusion modeling. This package uses a partial differential equation method and a simplex routine to obtain the parameters of the DDM, and uses the calculated cumulative density function (CDF) of the predicted RTs to estimate the goodness of fit using a Kolmogorov-Smirnov (KS) function (Voss and Voss, 2008 Voss et al., 2015). The DDM was fit separately for each task (lane change/image discrimination task) and each participant. We also calculated R 2 values as an additional measure to examine the goodness of fit of the model.

Data Analysis

Only the correct trials were used for the RT analysis. In the dual-task conditions, if the response to both tasks was correct, that trial was included in the analysis. The trials in which the reaction time to each of the tasks was 𼈀 ms and ϡ,500 ms were excluded from the analysis (3.48% of the trials). To quantify the effect of SOA on RTs and DDM parameters, one-way repeated-measures ANOVAs were used and to quantify the effect of SOA and task conditions on RTs, accuracies, and DDM parameters, two-way repeated-measures ANOVAs were used. A Greenhouse-Geisser correction was performed when sphericity had been violated. To compare the threshold, slope, and shift of the logistic regression function between the two task conditions, a paired t-test was used. We also performed three-way repeated measure ANOVAs with task condition, task order and SOA as three factors. The details of the statistical results are placed in Supplementary Tables S1–S3. In addition, we used t-test to statistically compare RTs, accuracies and DDM parameters between task conditions (dual vs. single/predictable vs. unpredictable) for each SOA. The details of the statistical tests for this analysis are placed in Supplementary Tables S7–S9. False Discovery Rate correction (Benjamini and Hochberg, 1995) was applied in all cases that multiple comparisons were performed.

We used a logistic regression model to examine the effect of SOA and OP on the order of the response of the two tasks. The probability that the lane change response was initiated before the image discrimination response was determined by the following formula:

where P stands for the probability that the lane change task was responded to first and C stands for SOAs. Parameters β0 and β1 were calculated for each participant. The model was fit separately on the data from the two dual-task conditions. A maximum likelihood estimation procedure was used for curve fitting.


Neural correlates of change detection and change blindness

Functional magnetic resonance imaging (fMRI) of subjects attempting to detect a visual change occurring during a screen flicker was used to distinguish the neural correlates of change detection from those of change blindness. Change detection resulted in enhanced activity in the parietal and right dorsolateral prefrontal cortex as well as category-selective regions of the extrastriate visual cortex (for example, fusiform gyrus for changing faces). Although change blindness resulted in some extrastriate activity, the dorsal activations were clearly absent. These results demonstrate the importance of parietal and dorsolateral frontal activations for conscious detection of changes in properties coded in the ventral visual pathway, and thus suggest a key involvement of dorsal–ventral interactions in visual awareness.


Attention Theory

The relation between attention and memory has been studied extensively for over a century (see Mulligan, 2008 for a review). Attention is typically viewed as a focusing process that plays a critical role in encoding, short-term memory, and long-term memory (see Jonides et al., 2008, for a review). The focus of attention can be placed on encoding, which is a perceptual process, as well as memories stored from past experiences. Information that is encoded and comes into the focus of attention can displace other memory contents from attentional focus. Attentional processes control what information enters into the focused state. This on-line processing explains the limited storage capacity of short-term memory. Attentional processes not only serve as “gating” mechanisms and determine what enters consciousness, but also how the information is maintained, and how additional information is retrieved from short-term and long-term memory storage. Memory performance degradation is typically explained in terms of either the decay of memory traces (engrams) in cortical structures (Loprinzi et al., 2017) or the interference among items held in perceptual memory, short-term, and long-term memory (Jonides et al., 2005). Of central importance to the present review is how concurrent physical activity may moderate the ability to overcome interference.

Attention Research: Experimental Evidence

A series of experiments examined the effects of walking on young and older adults’ memory. The studies were designed to test predictions derived from the Selection, Optimization, and Compensation (SOC) theory (Baltes and Baltes, 1990 Baltes and Lindenberger, 1997), which focuses on age-related shifts in the quality of sensorimotor processing. The SOC theory posits that individuals respond to environmental challenges via selection and modification of task goals and optimization of goal-directed compensatory strategies required to achieve those goals. With age-related declines in sensory acuity and proprioception, older adults are predicted to allocate increasingly more attentional resources to maintaining desirable levels of balance and walking control. Lindenberger et al. (2000) employed a dual-task methodology to assess the magnitude of dual-task interference between memorization of words and walking. 47 young (ages 20� years), 45 middle-aged (40�), and 48 old (60�) adults were trained to walk quickly and accurately on narrow paths that differed in movement complexity. In separate sessions, participants encoded 16 words presented auditorily while sitting, standing, or walking for 2.8 min on either track. Immediate serial recall tests were administered. Recall performance in the context of walking was lower than recall performance in the context of seated and standing conditions. Middle- and older-age participants showed a 22% loss in serial recall under a simple walk condition and a 36% loss during the complex walk whereas young adults showed no loss following the simple walk and a 19% loss during the complex walk. Similar findings were obtained from a subsequent study (Li et al., 2001) which retained the methods employed by Lindenberger et al. (2000) but individualized the walking and memory demands for each participant. Compared to a seated condition, older adults (60� year) showed significantly poorer serial recall performance following walking than did young adults (20� year). Younger adults showed no dual-task interference effect when performing a simple walk condition, and interference only when the walking task was more challenging (e.g., path obstacles).

A series of experiments conducted by Helton and colleagues focused on dual-task challenges presented in naturalistic conditions. In one study (Epling et al., 2016), young adults were asked to remember 20 words presented auditorily at 15-s intervals during a vigorous 5-min run on an outdoor running track. Immediately following the run, participants completed a 90-s written word-recall test. Significantly fewer words were recalled following the run than when the participant encoded words while seated. Similar results were obtained in two studies that evaluated dual-task interference when encoding words while performing traverse wall climbing (bouldering). Green and Helton (2011) assessed dual-task costs of bouldering across a climbing wall for three minutes and encoding 20 words presented auditorily every 8 s with a 14-s pause after the final word. Immediately following the climb, participants completed a 90-s written word-recall test. Compared to a single-task condition in which encoding occurred while seated, the dual-task condition resulted in a 50% decrease in recall performance. In a replication study, Darling and Helton (2014) presented words with irregular timing and the duration of traverse climbing was increased to 5 min. Compared to word recall performance in a single non-exercise condition, participants recalled nearly 40% fewer words. Results of these two studies suggested to the researchers that effortful processing required to traverse the wall climb interfered with the rehearsal and maintenance of words to be recalled.

Summary of Attentional Allocation Research

All studies reviewed here that were conducted on the basis of hypotheses drawn from attention theories consistently show evidence of CMI that negatively impacted the encoding of declarative information into long-term memory. The magnitude of interference appears to be age-related, with middle-age, and older adults showing less effective word memorization than younger adults. The data suggest that dual-task conditions that require motor movement planning and corrections compete with the ability to retrieve strategies from memory storage that are required for processing cognitive tasks, and thus negatively affect the encoding of semantic information into long-term memory storage.


Contents

Loftus, Miller, and Burns conducted the original misinformation effect study in 1978. Participants were shown a series of slides, one of which featured a car stopping in front of a stop sign. After viewing the slides, participants read a description of what they saw. Some of the participants were given descriptions that contained misinformation, which stated that the car stopped at a yield sign. Following the slides and the reading of the description, participants were tested on what they saw. The results revealed that participants who were exposed to such misinformation were more likely to report seeing a yield sign than participants who were not misinformed. [7]

Similar methods continue to be used in misinformation effect studies. Standard methods involve showing subjects an event, usually in the form of a slideshow or video. The event is followed by a time delay and introduction of post-event information. Finally, participants are retested on their memory of the original event. [8] The original study paved the way for multiple replications of the effect [ specify ] in order to test things such as the specific processes cause the effect to occur in the first place and how individual differences influence susceptibility to the effect.

Functional magnetic resonance imaging (fMRI) from 2010 pointed to certain brain areas which were especially active when false memories were retrieved. Participants studied photos during an fMRI. Later, they viewed sentences describing the photographs, some of which contained information conflicting with the photographs. One day later, participants returned for a surprise item memory recognition test on the content of the photographs. Results showed that some participants created false memories, reporting the verbal misinformation conflicting with the photographs. [9] During the original event phase, increased activity in left the fusiform gyrus and the right temporal/occipital cortex was found which may have reflected the attention to visual detail, associated with later accurate memory for the critical item(s) and thus resulted in resistance to the effects of later misinformation. [9] Retrieval of true memories was associated with greater reactivation of sensory-specific cortices, for example, the occipital cortex for vision. [9] Electroencephalography research on this issue also suggests that the retrieval of false memories is associated with reduced attention and recollection related processing relative to true memories. [10]

It is important to note that not everyone is equally susceptible to the misinformation effect. Individual traits and qualities can either increase or decrease one's susceptibility to recalling misinformation. [7] Such traits and qualities include age, working memory capacity, personality traits and imagery abilities.

Age Edit

Several studies have focused on the influence of the misinformation effect on various age groups. [11] Young children — especially pre-school-aged children — are more susceptible than older children and adults to the misinformation effect. [12] [13] [11] Young children are particularly susceptible to this effect as it relates to peripheral memories and information, as some evidence suggests that the misinformation effect is stronger on an ancillary, existent memory than on a new, purely fabricated memory. This effect is redoubled if its source is in the form of a narrative rather than a question. [14] However, children are also more likely to accept misinformation when it is presented in specific questions rather than in open-ended questions. [12]

Additionally, there are different perspectives regarding the vulnerability of elderly adults to the misinformation effect. Some evidence suggests that elderly adults are more susceptible to the misinformation effect than younger adults. [11] [15] [13] Contrary to this perspective, however, other studies hold that older adults may make fewer mistakes when it comes to the misinformation effect than younger ones, depending on the type of question being asked and the skillsets required in the recall. [16] This contrasting perspective holds that the defining factor when it comes to age, at least in adults, depends largely on cognitive capacity, and the cognitive deterioration that commonly accompanies age to be the typical cause of the typically observed decline. [16] Additionally, there is some research to suggest that older adults and younger adults are equally susceptible to misinformation effects. [17]

Working Memory Capacity Edit

Individuals with greater working memory capacity are better able to establish a more coherent image of an original event. Participants performed a dual task: simultaneously remembering a word list and judging the accuracy of arithmetic statements. Participants who were more accurate on the dual task were less susceptible to the misinformation effect, which allowed them to reject the misinformation. [7] [18]

Personality Traits Edit

The Myers Briggs Type Indicator is one type of test used to assess participant personalities. Individuals were presented with the same misinformation procedure as that used in the original Loftus et al. study in 1978 (see above). The results were evaluated in regards to their personality type. Introvert-intuitive participants were more likely to accept both accurate and inaccurate post-event information than extrovert-sensate participants. Researchers suggested that this likely occurred because introverts are more likely to have lower confidence in their memory and are more likely to accept misinformation. [7] [19] Individual personality characteristics, including empathy, absorption and self-monitoring, have also been linked to greater susceptibility. [11] Furthermore, research indicates that people are more susceptible to misinformation when they are more cooperative, dependent on rewards, and self-directed and have lower levels of fear of negative evaluation. [13]

Imagery Abilities Edit

The misinformation effect has been examined in individuals with varying imagery abilities. Participants viewed a filmed event followed by descriptive statements of the events in a traditional three-stage misinformation paradigm. Participants with higher imagery abilities were more susceptible to the misinformation effect than those with lower abilities. The psychologists argued that participants with higher imagery abilities were more likely to form vivid images of the misleading information at encoding or at retrieval, therefore increasing susceptibility. [7] [20]

Paired Participants Edit

Some evidence suggests that participants, if paired together for discussion, tend to have a homogenizing effect on the memory of one another. In the laboratory, paired participants that discussed a topic containing misinformation tended to display some degree of memory blend, suggesting that the misinformation had diffused among them. [21]

Time Edit

Individuals may not be actively rehearsing the details of a given event after encoding, as psychologists have found that the likelihood of incorporating misinformation increases as the delay between the original event and post-event information increases. [8] Furthermore, studying the original event for longer periods of time leads to lower susceptibility to the misinformation effect, due to increased rehearsal time. [8] Elizabeth Loftus' discrepancy detection principle argue that people's recollections are more likely to change if they do not immediately detect discrepancies between misinformation and the original event. [11] [22] At times people recognize a discrepancy between their memory and what they are being told. [23] People might recollect, "I thought I saw a stop sign, but the new information mentions a yield sign, I guess I must be wrong, it was a yield sign." [23] Although the individual recognizes the information as conflicting with their own memories, they still adopt it as true. [11] If these discrepancies are not immediately detected they are more likely to be incorporated into memory. [11]

Source Reliability Edit

The more reliable the source of the post-event information, the more likely it is that participants will adopt the information into their memory. [8] For example, Dodd and Bradshaw (1980) used slides of a car accident for their original event. They then had misinformation delivered to half of the participants by an unreliable source: a lawyer representing the driver. The remaining participants were presented with misinformation, but given no indication of the source. The misinformation was rejected by those who received information from the unreliable source and adopted by the other group of subjects. [8]

Discussion and Rehearsal Edit

Psychologists have also evaluated whether discussion impacts the misinformation effect. One study examined the effects of discussion in groups on recognition. The experimenters used three different conditions: discussion in groups with a confederate providing misinformation, discussion in groups with no confederate, and a no-discussion condition. They found that participants in the confederate condition adopted the misinformation provided by the confederate. However, there was no difference between the no-confederate and no-discussion conditions, providing evidence that discussion (without misinformation) is neither harmful nor beneficial to memory accuracy. [24] Additionally, research has found that collaborative pairs showed a smaller misinformation effect than individuals, as collaborative recall allowed witnesses to dismiss misinformation generated by an inaccurate narrative. [25] Furthermore, there is some evidence suggesting that witnesses who talk with each other after watching two different videos of a burglary will claim to remember details shown in the video seen by the other witness. [26]

State of Mind Edit

Various inhibited states of mind such as drunkenness and hypnosis can increase misinformation effects. [11] Assefi and Garry (2002) found that participants who believed they had consumed alcohol showed results of the misinformation effect on recall tasks. [27] The same was true of participants under the influence of hypnosis. [28]

Arousal and Stress After Learning Edit

Arousal induced after learning reduces source confusion, allowing participants to better retrieve accurate details and reject misinformation. In a study of how to reduce the misinformation effect, participants viewed four short film clips, each followed by a retention test, which for some participants included misinformation. Afterward, participants viewed another film clip that was either arousing or neutral. One week later, the arousal group recognized significantly more details and endorsed significantly fewer misinformation items than the neutral group. [29] Similarly, research also suggests that inducing social stress after presenting misinformation makes individuals less likely to accept misinformation. [30]

Anticipation Edit

Educating participants about the misinformation effect can enable them to resist its influence. However, if warnings are given after the presentation of misinformation, they do not aid participants in discriminating between original and post-event information. [11]

Psychotropic Placebos Edit

Research published in 2008 showed that placebos enhanced memory performance. Participants were given a placebo "cognitive enhancing drug" called R273. When they participated in a misinformation effect experiment, people who took R273 were more resistant to the effects of misleading post-event information. [31] As a result of taking R273, people used stricter source monitoring and attributed their behavior to the placebo and not to themselves. [31]

Sleep Edit

Controversial perspectives exist regarding the effects of sleep on the misinformation effect. One school of thought supports the idea that sleep can increase individuals' vulnerability to the misinformation effect. In a study examining this, some evidence was found that misinformation susceptibility increases after a sleeping cycle. In this study, the participants that displayed the least degree of misinformation susceptibility were the ones who had not slept since exposure to the original information, indicating that a cycle of sleep increased susceptibility. [16] Researchers have also found that individuals display a stronger misinformation effect when they have a 12-hour sleep interval in between witnessing an event and learning misinformation than when they have a 12-hour wakefulness interval in between the event and the introduction of misinformation. [32]

In contrast, a different school of thought holds that sleep deprivation leads to greater vulnerability to the misinformation effect. This view holds that sleep deprivation increases individual suggestibility. [33] This theory posits that this increased susceptibility would result in an related increase in the development of false memories. [21] [34]

Other Edit

Most obviously, leading questions and narrative accounts can change episodic memories and thereby affect witness' responses to questions about the original event. Additionally, witnesses are more likely to be swayed by misinformation when they are suffering from alcohol withdrawal [25] [35] or sleep deprivation, [25] [36] when interviewers are firm as opposed to friendly, [25] [37] and when participants experience repeated questioning about the event. [25] [38]

The misinformation effect can have dire consequences on decision making that can have harmful personal and public outcomes in a variety of circumstances. For this reason, various researchers have participated in the pursuit of a means to counter its effects, and many models have been proposed. As with Source Misattribution, attempts to unroot misinformation can have lingering unaddressed effects that do not display in short term examination. Although various perspectives have been proposed, all suffer from a similar lack of metanalytic examination.

False Confirmation Edit

One of the problems with countering the misinformation effect, linked with the complexity of human memory, is the influence of information, whether legitimate or falsified, that appears to support the false information. The presence of these confirmatory messages can serve to validate the Misinformation as presented, making it more difficult to unroot the problem. This is particularly present in situations where the person has a desire for the information to be legitimate. [39]

Directly Oppositional Messages Edit

A common method of unrooting false concepts is presenting a contrasting, “factual” message. While this would intuitively be a good means of portraying the information to be inaccurate, this type of direct opposition has been linked to an increase in misinformation belief. Some researchers hypothesize that the counter message must have at least as much support, if not more, than the initial message to present a fully developed countermodel for consideration. Otherwise, the recipient may not remember what was wrong about the information and fall back on their prior belief model due to lack of support for the new model. [40]

Exposure to the Original Source Edit

Some studies suggest that the misinformation effect can occur despite exposure to accurate information. [41] This effect has been demonstrated when the participants have the ability to access an original, accurate video source at whim, and has even been demonstrated when the video is cued to the precise point in time where video evidence that refutes the misinformation is present. [41] Written and photographic contradictory evidence have also been shown to be similarly ineffective. Ultimately, this demonstrates that exposure to the original source is still not guaranteed to overcome the misinformation effect. [41]

There are a few existing evidence-based models for addressing the misinformation effect. Each of these, however, have their own limitations that impact their effectiveness.

Increased Self Regard Edit

Some evidence has been shown to suggest that those suffering from the misinformation effect can often tell they are reporting inaccurate information but are insufficiently confident in their own recollections to act on this impression. [42] As such, some research suggests that increased self-confidence, such as in the form of self-affirmative messages and positive feedback, can weaken the misinformation effect. [42] Unfortunately, due to the difficulty of introducing increased self-regard in the moment, these treatment methods are held to not be particularly realistic for use in a given moment. [42]

Pretesting as a Means of Preventing the Misinformation Effect Edit

Another direction of study in preventing the misinformation effect is the idea of using a pretest to prevent the misinformation effect. This theory posits that a test, applied prior to the introduction of misleading information, can help maintain the accuracy of the memories developed after that point. [43] This model, however, has two primary limitations: its effects only seem to hold for one item at a time, and data supports the idea that it increases the impact of the information on the subsequent point of data. Pretesting also, paradoxically, has been linked with a decrease in accurate attributions from the original sample. [43]

The Use of Questions Edit

Another model with some support is that of the use of questions. This model holds that the use of questions rather than declaratory statements prevents the misinformation effect from developing, even when the same information is presented in both scenarios. In fact, the use of questions in presenting information after the fact was linked with increased correct recall, and further with an increase in perfect recall among participants. The advocates of this view hold that this occurs because the mind incorporates definitive statements into itself, whereas it does not integrate questions as easily. [44]

Post-Misinformation Corrections and Warnings Edit

Correcting misinformation after it has been presented has been shown to be effective at significantly reducing the misinformation effect. [45] Similarly, researchers have also examined whether warning people that they might have been exposed to misinformation after the fact impacts the misinformation effect. [46] [11] A meta-analysis of studies researching the effect of warnings after the introduction of misinformation found that warning participants about misinformation was an effective way to reduce — though not eliminate — the misinformation effect. [46] However, the efficacy of post-warnings appears to be significantly lower when using a recall test. [46] Warnings also appear to be less effective when people have been exposed to misinformation more frequently. [11]

Current research on the misinformation effect presents numerous implications for our understanding of human memory overall.

Variability Edit

Some reject the notion that misinformation always causes impairment of original memories. [11] Modified tests can be used to examine the issue of long-term memory impairment. [11] In one example of such a test,(1985) participants were shown a burglar with a hammer. [47] Standard post-event information claimed the weapon was a screwdriver and participants were likely to choose the screwdriver rather than the hammer as correct. In the modified test condition, post-event information was not limited to one item, instead participants had the option of the hammer and another tool (a wrench, for example). In this condition, participants generally chose the hammer, showing that there was no memory impairment. [47]

Rich False Memories Edit

Rich false memories are researchers' attempts to plant entire memories of events which never happened in participants' memories. Examples of such memories include fabricated stories about participants getting lost in the supermarket or shopping mall as children. Researchers often rely on suggestive interviews and the power of suggestion from family members, known as “familial informant false narrative procedure.” [11] Around 30% of subjects have gone on to produce either partial or complete false memories in these studies. [11] There is a concern that real memories and experiences may be surfacing as a result of prodding and interviews. To deal with this concern, many researchers switched to implausible memory scenarios. [11] Researchers have also found that they were able to induce rich false memories of committing a crime in early adolescence using a false narrative paradigm. [48]

Daily Applications: Eyewitness Testimony Edit

The misinformation effect can be observed in many situations. In particular, research on the misinformation effect has frequently applied to eyewitness testimony and has been used to evaluate the trustworthiness of eyewitnesses’ memory. [4] [13] [6] After witnessing a crime or accident there may be opportunities for witnesses to interact and share information. [4] [6] Late-arriving bystanders or members of the media may ask witnesses to recall the event before law enforcement or legal representatives have the opportunity to interview them. [25] Collaborative recall may lead to a more accurate account of what happened, as opposed to individual responses that may contain more untruths after the fact. [25] However, there have also been instances where multiple eyewitnesses have all remembered information incorrectly. [13] Remembering even small details can be extremely important for eyewitnesses: A jury's perception of a defendant's guilt or innocence could depend on such a detail. [2] If a witness remembers a moustache or a weapon when there was none, the wrong person may be wrongly convicted. [3]


Materials and Methods

Participants

Twenty healthy, right-handed adults (11 females), aged 20�, participated in the study. All participants had normal or corrected to normal vision. Additionally, all participants were not expert video game players, as defined by having less than 2 h of video-game usage per month in the past 2 years. All participants gave informed consent and were compensated for their participation.

Stimuli and Procedure

The dual-task paradigm consisted of a lane change driving task and an image discrimination task. The driving environment was designed in the Unity 3D game engine. Participants sat at a distance of 50 cm from a 22″ LG monitor with a refresh rate of 60 Hz and a resolution of 1,920 × 1,080 and responded to the tasks using a computer keyboard.

The driving environment consisted of a three-lane, desert road, without left/right turns or inclining/declining hills. Driving stimuli, composed of two rows of traffic cones (three cones in each row Figure 2A), were presented on the two sides of one of the lanes in each trial, and the participants had to immediately redirect the car to the lane with the cones and pass through the cones. The space between the two rows of cones was such that the car could easily pass through them without collision. The cones were always presented in the lanes immediately to the left or immediately to the right of the car’s lane so that the participants had to change only one lane per trial. The lane change was done gradually: the participant had to hold the corresponding key to direct the car in between the two rows of cones, and then release the key when the car was situated correctly. Any early or late key press or release would cause a collision with the cones and a performance loss in that trial. The fixation cross was jittered for 100 ms to provide online feedback in case of a collision with the traffic cones. The participants were instructed not to change lane before the cones appeared. Trials in which participants changed lane before the presentation of the cones were considered false and removed from the analysis. Using this method, we could divide a continuous driving task into individual trials with predetermined onset and ends. At the beginning of the block, participants speeded up to 80 km/h using the “up” arrow key with the middle finger of the right hand. During the block, the speed was kept constant, and the lane change was performed by pressing the right and left arrow with the middle and index fingers of their right hand, respectively. For the image discrimination task, a single image of either a scene or a face was presented for 150 milliseconds centered at 2° eccentricity above the fixation cross (Figure 2B). The size of the image was 2.5° of visual angle. Participants pressed the “x” and “z” keys on the computer keyboard with the middle and index fingers of their left hand to determine whether the image was a face or a scene, respectively. The images were pseudo-randomly selected from a set of 864 images of scenes and 435 images of faces. We selected only natural scenes and neutral faces. If participants responded incorrectly, the green fixation cross turned red, and if they responded late, it turned orange for 100 ms. The length of each trial was 3 s, and the inter-trial interval varied randomly from 0.5 to 1.5 s. For the first trial in each block, the onset of the trial was set to 2 s after the beginning of the block. The end of the trial was set to when the rear end of the car reached the end of the set of traffic cones.

Figure 2. Dual-task paradigm. (A) A sample display showing the driving stimulus consisting of two rows of traffic cones in the middle driving lane. The cones were randomly presented in each lane, and participants had to drive through them without collision. (B) A sample display showing an image discrimination presented above the fixation point. Participants determined if the image was a face or a scene. (C) The sequence of events for a sample trial in which the image task was presented first (left), and another in which the driving task was presented first (right). The inter-trial interval (ITI) varied between 0.5 and 1.5 s. The image lasted for 150 ms, and the cones were presented 30, 100, 300, or 600 before or after the image. Participants had to perform a lane change immediately after the appearance of the cones, and an image discrimination task immediately after the presentation of the image.

The experiment consisted of two different conditions: (1) “Predictable” task order condition, and (2) “Unpredictable” task order condition. In two experimental conditions, the two tasks were presented with eight possible SOAs (�, �, �, �, +30, +100, +300 and +600 ms). In the negative SOAs, the image discrimination was presented first (image-first, Figure 2C), and in the positive SOAs, the lane change was presented first (lane change-first, Figure 2C). In the Predictable conditions, the order of the presentation was fixed, so that in two of the four blocks, the driving task was presented first, and in the other two, the image discrimination task was presented first. In the Unpredictable condition, the order of the presentation of the two tasks was not predictable in each trial. Trials with driving as the first task were interleaved with trials with the image discrimination as the first task. Before the start of each block, participants were informed about the type of the block.

In addition to the dual-task conditions, participants performed two single-task conditions: (1) single driving task and (2) single image discrimination task. In the single-task conditions, both the lane change and image stimuli were presented, but the participant only responded to one of them, ignoring the other. In the single image discrimination condition, the driving was on autopilot, and participants only responded to the images. In the single lane change condition, participants performed the lane change task and ignored the images.

Participants were told to focus on the fixation cross at the center of the page and respond to each task as fast as possible. At the end of each block, participants were informed about their performance on each task as well as their total performance. The performance in the driving task was calculated as the percentage of trials in which the participant passed through the cones without collision. The performance in the image discrimination task was calculated as the percentage of correct identifications.

Participants completed four blocks of 64 trials for each dual-task condition and two blocks of 32 trials for each single-task condition. There was a 1-min interval between blocks and a 5-min break after finishing all the blocks in each condition. The order of the blocks was counterbalanced across participants.

Before performing the main experiment, all participants performed a block of 20 trials for every single-task. If their accuracy was 80% or higher, they proceeded to the main experimental blocks. Otherwise, they repeated blocks of 40 trials for each task until they reached 80% accuracy. After the single-task training, participants performed the dual-task training block. The dual-task training was similar to the single-task training block, with the difference that if after 20 trials, the dual-task performance did not reach the 75% threshold, the training was repeated with blocks of 50 trials.

Drift Diffusion Model Fitting

To investigate if the two tasks were processed serially, or in parallel we used a DDM in which each trial was modeled as a combination of a non-decision time and a decision time consisting of a random drift towards decision bound (Figure 1). Model parameters consisted of: (1) parameter z denoting the starting point of the decision process, (2) parameter a denoting the decision threshold, (3) parameter v representing the speed of information accumulation or drift rate, and (4) parameter t0 denoting the non-decision time pertaining to the combination of all other times in the trial excluding the drift-diffusion time. The DDM was implemented in the current study, by fitting the parameters z, a, v, and t0. We modified the DDM, so that z and a were independent of SOA, and v and t0 were dependent on SOA. Therefore, in the modified DDM, four values were fit for the parameter v and four values for the parameter t0 corresponding to the four SOAs, one value for the parameter a and one value for the parameter z across all SOAs.

We used the Fast-dm package, developed by Voss and Voss (2007), for model fitting. Fast-dm is a package for fast drift-diffusion modeling. This package uses a partial differential equation method and a simplex routine to obtain the parameters of the DDM, and uses the calculated cumulative density function (CDF) of the predicted RTs to estimate the goodness of fit using a Kolmogorov-Smirnov (KS) function (Voss and Voss, 2008 Voss et al., 2015). The DDM was fit separately for each task (lane change/image discrimination task) and each participant. We also calculated R 2 values as an additional measure to examine the goodness of fit of the model.

Data Analysis

Only the correct trials were used for the RT analysis. In the dual-task conditions, if the response to both tasks was correct, that trial was included in the analysis. The trials in which the reaction time to each of the tasks was 𼈀 ms and ϡ,500 ms were excluded from the analysis (3.48% of the trials). To quantify the effect of SOA on RTs and DDM parameters, one-way repeated-measures ANOVAs were used and to quantify the effect of SOA and task conditions on RTs, accuracies, and DDM parameters, two-way repeated-measures ANOVAs were used. A Greenhouse-Geisser correction was performed when sphericity had been violated. To compare the threshold, slope, and shift of the logistic regression function between the two task conditions, a paired t-test was used. We also performed three-way repeated measure ANOVAs with task condition, task order and SOA as three factors. The details of the statistical results are placed in Supplementary Tables S1–S3. In addition, we used t-test to statistically compare RTs, accuracies and DDM parameters between task conditions (dual vs. single/predictable vs. unpredictable) for each SOA. The details of the statistical tests for this analysis are placed in Supplementary Tables S7–S9. False Discovery Rate correction (Benjamini and Hochberg, 1995) was applied in all cases that multiple comparisons were performed.

We used a logistic regression model to examine the effect of SOA and OP on the order of the response of the two tasks. The probability that the lane change response was initiated before the image discrimination response was determined by the following formula:

where P stands for the probability that the lane change task was responded to first and C stands for SOAs. Parameters β0 and β1 were calculated for each participant. The model was fit separately on the data from the two dual-task conditions. A maximum likelihood estimation procedure was used for curve fitting.


Frontiers in Psychology

The editor and reviewers' affiliations are the latest provided on their Loop research profiles and may not reflect their situation at the time of review.


  • Download Article
    • Download PDF
    • ReadCube
    • EPUB
    • XML (NLM)
    • Supplementary
      Material
    • EndNote
    • Reference Manager
    • Simple TEXT file
    • BibTex


    SHARE ON

    GENERAL DISCUSSION

    By having participants perform a tracking task while conversing with a friend or a confederate, we were able to record fine-grained, dynamic changes in attention demands during conversation. The two experiments we reported have shown that the attentional demands associated with natural conversation change rapidly. Specifically, we observed that during talking segments, tracking performance declined at the beginning and improved at the end and that during listening segments, performance improved at the beginning and declined at the end. To the best of our knowledge, this is the first demonstration of the dynamical changes in attentional demands during different stages of conversation at a millisecond time scale.

    The specific pattern of these dynamical changes indicates that speech planning is the most resource-demanding aspect of conversation. This is further supported by the similarity between tracking task performance during Pause and Talk segments, and by the gradual decline in tracking performance that occurs during Prepare segments. These show that the detrimental effects of talking actually occur even before and without any acoustic instantiation of speech. Indeed, our findings indicate that the attentional detriment associated with speech planning is spread across talking, preparing, and listening segments. This has some important implications for theories of language comprehension and production, which we will review in more detail below.

    We also found that increasing the difficulty of the tracking task exaggerated the interference from conversation. Because increasing tracking task difficulty increases monitoring requirements, this finding indicates that the cognitive demands of speech monitoring overlap with those of speech planning, and therefore that the two processes tap shared resources. This provides useful constraints for applying theories of dual-task performance (e.g., Wickens, 2008) to multitasking during conversation. This is also reviewed in more detail below.

    Finally, we believe that the approach of recording continuous measures of secondary task performance will be a valuable tool for other researchers interested in studying how attention shifts across multitasking scenarios. Our results indicate that this approach is most profitably used in conjunction with an advanced statistical analysis of dynamical change, such as the GCAs that we employed in this work.

    We note that although one of our findings, that talking is more attention demanding than listening, has been reported before, previous research has only shown this for non-naturalistic linguistic tasks that either involve no true linguistic communication (Kunar et al., 2008) or at best, involve an interaction with a pre-recorded script (Almor, 2008). Indeed, our fine-grained temporal analyses allowed us to explain why talking is more demanding than listening. Specifically, we claim that talking involves simultaneous planning and monitoring, both of which require considerable attentional resources.

    We now turn to discuss in detail the theoretical implications of our findings for theories of language comprehension and production, and for dual-task research.

    Language Comprehension and Production during Dual-Task Scenarios

    The vast differences we have shown between tracking performance during listening and talking have important implications for models of language comprehension and production. For comprehension, our results support theories that argue for flexibility in the effort allocated to comprehension. One example is the good enough processing theory of sentence comprehension (Christianson et al., 2006 Ferreira, Bailey, & Ferraro, 2002 Ferreira & Patson, 2007), which states that comprehenders process linguistic information only to the extent necessary for deriving a possible interpretation. According to this approach, full syntactic parsing of the input is effortful and only occurs in special circumstances, such as when the derived interpretation is deemed wrong due to inconsistency with preceding context or upon receiving successive input. Another example is the minimalist approach to discourse processing proposed by McKoon and colleagues (McKoon & Ratcliff, 1990 McKoon & Ratcliff, 1989a McKoon & Ratcliff, 1989b). According to this approach, discourse relations such as inferences and co-reference are not always computed by comprehenders. Whether or not such relations are processed depends on the comprehender’s goals (Greene, McKoon & Ratcliff, 1992) and level of engagement (Love & McKoon, 2011) as well as on processing difficulty. The main tenet of both theories is that the depth of processing during comprehension varies as a function of the effort required for processing, the available resources, and the goals of the comprehender. While listening, when resources are required for a secondary task, for example in the difficult conditions of target tracking, interlocutors can allocate resources to the secondary task. This was indicated in participants’ tracking performance at the beginning and end of listening segments. Note that according to these theories, fewer resources do not entail reduced comprehension but simply that successful comprehension can be achieved with fewer resources and that the incentive to divert resources away from comprehension varies according to the factors we discussed.

    In contrast to comprehension, production is much more effortful, requiring precise planning, lexical access, syntactic formulation, motor execution, and monitoring at all levels (Postma, 2000). Even the monitoring aspect of production (which could be argued to be a similar mechanism as that employed during comprehension) appears to consume more resources during production than comprehension, likely because every stage of linguistic production requires the use of the monitor or separate monitors, whereas a conversational partner’s linguistic input may not be scrutinized with such rigor. Other authors have similarly argued (as we have) that the planning and monitoring stages of speech production pose the greatest demands on attention (e.g., Blackmer & Mitton, 1991 Ferreira & Pashler, 2002 Oomen & Postma, 2002). Our data show the progress from only speech planning, to the more demanding combination of ongoing speech and planning, and finally to the most attentionally demanding mixture of speech, planning, and monitoring. As more processes are combined, the demands on attention increase, leading to worse performance on the concurrent tracking task. One implication of this interpretation is that speech planning itself is not what leads to worse performance but that the detriment is a result of planning overlapping with other processes. This is mirrored in our results which showed that the difference between listening and speaking varied as a function of the difficulty of the visual task. When the visual task was more demanding, more attention was gradually directed towards the visual task during listening and the opposite was true during speaking.

    There were some intriguing differences between our two experiments. The main difference is that generally E2 appeared to be easier: participants from E2 said the conversations were more natural and typical than the participants from E1 there was a main effect of Conversation on the tracking task in E1 but not in E2 and GCA graphs show that the effect of talking was more exaggerated in E1 than in E2. However, we must tread carefully when interpreting these apparent differences. For one thing, E1 only used one confederate. Thus, the results may only be applied to situations in which someone is talking with a lab confederate or possibly even just the individual we used in our experiment. E2, on the other hand, shows reliable differences in tracking task performance when participants are conversing with different people: their friends. Previous research (Savitsky et al., 2011) has shown that friends are not as vigilant at tracking one another’s common ground (Clark & Brennan, 1991) as they are when speaking with a stranger. For example, participants tended to make more egocentric mistakes when following directions given by a friend than when following directions from a stranger. Savitsky et al. interpreted their findings as demonstrating an overestimation of the success of communicating with a close friend, something they term the closeness-communication bias. However, another possible interpretation is that intimates feel less impetus to invest the cognitive resources necessary for audience design when communicating with a friend because they do not attach high cost to temporary communication failures and corrections. This means that interlocutors may choose to invest more or less resources in the communication depending on their level of intimacy or perhaps on their overall level of comfort. While Savitsky’s study shows that this bias occurs in discourse-level comprehension, we expect it would easily translate into production as well. If participants, in speaking with friends, do not allocate as many resources to monitoring their speech (and discourse content especially) as they do when speaking with strangers, we would expect conversations with friends to be less detrimental to a concurrent task than conversations with strangers (or in our case conversations with a confederate). If indeed this closeness-communication bias did manifest during speech production, we might expect it to look similar to the observed differences between Prepare and Talk with Listen that we saw across our experiments. However, because the design of our study does not permit us to directly test these claims, disentangling these issues remains open for future research.

    Dual-Task Attention Model

    Wickens’s (2008) model of dual-task attention provides a useful framework for interpreting the results from our experiment. According to that model, detriments to the concurrent visuomotor task should be most apparent when the two tasks match along three possible dimensions: stage of processing, code of processing, and modality. This is indeed the case in the present study, as can be seen in the differences between talking and listening. We had earlier argued that talking involves planning, response, and monitoring, while listening involves monitoring (and this only minimally). Both planning and monitoring increase attentional load along the cognitive processing stage, while the mechanical aspects of speech tax the response processing stage. In our experiments, we observed increasing interference during planning and then a continued increase in interference with the addition of speaking and monitoring. Listening, on the other hand, only makes use of the cognitive processing stage. In other words, listening, which overlapped the least with the tracking task in terms of Wickens’s model’s processing dimensions, interfered with tracking the least.


    Watch the video: Stop Signal Task - Cognitive Psychology Paradigms and Task (June 2022).


Comments:

  1. De

    It is a pity, that I can not participate in discussion now. I do not own the necessary information. But this theme me very much interests.

  2. Barra

    At the risk of sounding like a layman, but still I will ask, where did this come from and who wrote it at all?

  3. Wendlesora

    Between us, in my opinion, this is obvious. I recommend looking for the answer to your question on google.com



Write a message