Date Published: September 19, 2019
Publisher: Public Library of Science
Author(s): Kenneth Hugdahl, Katarzyna Kazimierczak, Justyna Beresniewicz, Kristiina Kompus, Rene Westerhausen, Lars Ersland, Renate Grüner, Karsten Specht, Andreas Hahn.
Using fMRI, Hugdahl et al. (2015) reported the existence of a general-domain cortical network during active task-processing which was non-specific to the cognitive task being processed. They labelled this network the extrinsic mode network (EMN). The EMN would be predicted to be negatively, or anti-correlated with the classic default mode network (DMN), typically observed during periods of rest, such that while the EMN should be down-regulated and the DMN up-regulated in the absence of demands for task-processing, the reverse should occur when demands change from resting to task-processing. This would require alternating periods of task-processing and resting and analyzing data continuously when demands change from active to passive periods and vice versa. We were particularly interested in how the networks interact in the critical transition points between conditions. For this purpose, we used an auditory task with multiple cognitive demands in a standard fMRI block-design. Task-present (ON) blocks were alternated with an equal number of task-absent, or rest (OFF) blocks to capture network dynamics across time and changing environmental demands. To achieve this, we specified the onset of each block, and used a finite-impulse response function (FIR) as basis function for estimation of the fMRI-BOLD response. During active (ON) blocks, the results showed an initial rapid onset of activity in the EMN network, which remained throughout the period, and faded away during the first scan of the OFF-block. During OFF blocks, activity in the DMN network showed an initial time-lag where neither the EMN nor the DMN was active, after which the DMN was up-regulated. Studying network dynamics in alternating passive and active periods may provide new insights into brain network interaction and regulation.
Applying an inclusive conjunction analysis to fMRI data from nine different studies with a total of 187 subjects, and comprising nine different cognitive tasks, Hugdahl et al.  found a generalized cortical network that was independent of the specifics of the individual task and cognitive processes. The authors labeled this the extrinsic mode network (EMN), as a task non-specific network, with a fronto-temporo-parietal distribution, including the inferior and middle frontal gyri, inferior parietal lobule, supplementary motor area, and the inferior temporal gyrus. The EMN thus has a spatial architecture overlapping with what Fedorenko et al.  labeled the “cognitive flexibility network”, and Duncan  labeled the “multiple demand network” (see also ). Common for all these networks is that they are suggested to be general-domain networks, i.e. they show up-regulation of activity across a range of cognitive tasks, e.g. , , , . The general-domain networks further share characteristics with several domain-specific networks, such as the dorsal attention and salience networks , , , ), central executive network , , fronto-parietal network , and ventral attention network , see also Lee et al. , and Cabeza and Nyberg  for overviews. It is therefore clear that the dynamic interplay between various cortical networks, during different mind-states is far more complex than previously thought [19, 20, 21]. In addition, while there are similarities between the EMN and other task-positive networks with regard to overlapping activity patters, a clear difference is that the EMN is supposed to be a domain non-specific network, while the typical task-positive networks, like the dorsal attention and central executive networks are supposedly domain-specific networks. In the current study, we asked how the EMN as a general-domain network relates to the more known default mode network (DMN) , , , see also , , , , which is typically observed in the absence of specific tasks. It could be predicted that the DMN should be negatively, or anti- correlated with a general-domain network, like the EMN. This prediction is derived from previous findings that neuronal activity varies reciprocally between characteristic DMN- and task-positive network-areas, when studied under both resting-state , , , , and during active task conditions , , . Lustig et al.  used alternating blocks of passive viewing of a fixation cross and active processing of a semantic judgment task. Results showed increased activations in the left frontal cortex during task-processing and deactivation in the same area during passive fixation blocks. An opposite pattern was seen in the lateral parietal cortex, with deactivations during task processing. The study by Lustig et al.  therefore showed inverse activations and deactivations in brain areas linked to active task-processing compared to passive viewing. Following previous findings we therefore asked whether a similar relationship should hold for general-domain networks, and in particular for the EMN, when different cognitive tasks are alternated during the scanning session. Thus, we alternated brief periods with task presentations with brief periods of resting with no tasks present. Conventional analysis of block-design data is to subtract activity during OFF-blocks from activity during ON-blocks, where the OFF-blocks act as a baseline control-condition (we leave out here the discussion in the literature whether the assumption of “pure insertions” in block-designs is a valid assumption or not, cf. . The resulting activity pattern would thus reflect active task-processing. By subtracting activity obtained during task-processing ON-blocks from activity obtained during resting OFF-blocks, cf. , it should be possible to display activity that would be deactivated during task-processing blocks. We are not focusing here on whether task-positive and task-negative networks are anti-correlated per se in resting-state fMRI situations, cf. , , but on the time-dynamics of up- and down-regulations across the transitions between task-present and task-absent periods, with a focus on the interaction between the DMN and EMN networks. For this purpose, we used a finite-impulse response (FIR) function to model the BOLD response as implemented in the Statistical Parametric Mapping (SPM) analysis software (https://www.fil.ion.ucl.ac.uk/spm/). We re-analyzed fMRI data from >100 healthy individuals from a previous study in our laboratory , where the subjects had been tested with an auditory dichotic listening (DL) task with three instruction conditions that emphasized either perception , attention/vigilance  or executive control functions . We chose this task because it reflects the changing cognitive demands and coping situations during an ordinary day, including both low-, perception, and high-, executive control, level demands. The so called forced-attention DL paradigm was originally developed by Hugdahl and Andersson  for the study of the role of cognitive factors in auditory perception, and it has repeatedly been shown to produce valid and reliable results with regard to perception, attention and executive control, see e.g. , , , , , , , . We now report how the EMN interact with the DMN within a single paradigm which included alternating task-presence (ON) and task-absence (OFF) periods, and with recurring and varying cognitive tasks and demands. Such an experimental design would be a novel way of capturing the dynamic interaction between passive rest and active processing periods, going beyond a fMRI “resting period” data acquisition approach. In order to capture the dynamics of network up- and down-regulation over time, and especially at the transition points between ON- and OFF-blocks, we specified the onset of each condition and used a 5.5s finite-impulse response function (FIR) as basis function for estimation of the BOLD response. This approach modeled each scan per task-ON- and task-OFF-block separately, which would allow an analysis of the fine-grained dynamics in the critical time-window when the situation switched from active to passive, and from passive to active time periods.
The results of the averaged ON-OFF contrast revealed the typical task-positive activity pattern with bilateral activity in the auditory cortex and surroundings, and of the task-positive, EMN, network  in the prefrontal cortex, including the anterior and middle cingulate cortex, supplementary motor area (SMA/preSMA), and thalamus. The opposite contrast revealed areas, belonging to the task-negative, DMN, network , revealed significant activity in the precuneus, inferior parietal, and medial orbitofrontal areas, and in occipital areas (Figs 1 and 2) and (Tables 1 and 2). This could be activity returning to baseline during resting periods, as previously found, see e.g. , or as a new finding with increases above baseline in situations with alternating task-negative and task-positive periods.
To sum up the main findings, the ON- and OFF-blocks, corresponding to active task-processing versus passive resting, produced two orthogonal, non-overlapping activity patterns (see Figs 1 and 2). As is obvious from the figures, while task-present epochs resulted in a more anterior activity pattern, with the SMA/preSMA and the auditory cortices as the dominant activity regions, the task-absent epochs resulted in a more posterior activity pattern, including activity in the precuneus, and the parietal lobules as the dominant regions. These activity patterns would partially correspond to the EMN and DMM networks, respectively  . The FIR analysis of the network dynamics based on time-derivative contrasts in addition revealed that the EMN showed a relatively sharp onset of the up-regulation at the beginning of the ON-blocks, while there was a more gradual up-regulation of the DMN during OFF-blocks. As expected, the DMN was more active during the OFF-blocks (see Fig 3), but no abrupt change was seen in the time-derivative analysis at the ON-OFF block intersections (see Fig 4). In their meta-analysis of nine PET blood-flow studies  found significant decreases for the active minus passive task condition in cortical areas that today would qualify as the “classic” areas for the default mode network. One interpretation of the decreases in the Schulman et al.  study is that task-processing may inhibit activity in areas that otherwise would be tonically activated in the absence of a task, as was suggested by Popa et al. . These authors used electrophysiology recordings and found that local field potential power was lower in anterior cingulum and retrosplenial cortex during task-OFF compared to task-ON periods, while the reverse was found in somato-sensory association cortex and middle temporal gyrus. A further confirmation of this hypothesis would be if activity in approximately the same areas would be increased during passive resting periods in between active task processing periods, which the current results have shown. The areas showing decreased activity during active minus passive epochs in the Schulman et al.  study were primarily in the posterior cingulate/precuneus, dorsolateral and inferior frontal cortex, and in the inferior temporal gyrus (passive here meaning being exposed to the same stimulus, but without instruction to act on the stimulus). Approximately the same areas were activated in the present study, but now when subtracting active task-processing periods from activity during a resting period, which would confirm the hypothesis of inhibition of tonically active areas during phasic task-processing. Similarly, Schulman et al.  found increases in the visual cortex in the occipital lobe during active visual task processing, after averaging data from 10 different studies with visual tasks. This is paralleled in the present study which found corresponding increases in the auditory cortex in the temporal lobe to an auditory task. Previous studies have shown that the DMN is still up-regulated during task-presence periods but attenuated or suspended compared to task-related activity, e.g. , , . Although the present findings are in line with these results, we cannot say if the corresponding networks were down-regulated or merely attenuated during reversals. An advantage with the present paradigm over previous paradigms that have been used, e.g. , ,  is that the three different cognitive tasks were all embedded within the same experimental paradigm. The dichotic listening (DL) task is moreover exceptionally easy to understand and perform, so that the understanding of the task in itself does not require the allocation of additional cognitive resources which could confound task processing.