Multimodal network dynamics underpinning working memory

Complex human cognition arises from the integrated processing of multiple brain systems. However, little is known about how brain systems and their interactions might relate to, or perhaps even explain, human cognitive capacities. Here, we address this gap in knowledge by proposing a mechanistic framework linking frontoparietal system activity, default mode system activity, and the interactions between them, with individual differences in working memory capacity. We show that working memory performance depends on the strength of functional interactions between the frontoparietal and default mode systems. We find that this strength is modulated by the activation of two newly described brain regions, and demonstrate that the functional role of these systems is underpinned by structural white matter. Broadly, our study presents a holistic account of how regional activity, functional connections, and structural linkages together support integrative processing across brain systems in order for the brain to execute a complex cognitive process.


Statistics
For all statistical analyses, confirm that the following items are present in the figure legend, table legend, main text, or Methods section.

n/a Confirmed
The exact sample size (n) for each experimental group/condition, given as a discrete number and unit of measurement A statement on whether measurements were taken from distinct samples or whether the same sample was measured repeatedly The statistical test(s) used AND whether they are one-or two-sided Only common tests should be described solely by name; describe more complex techniques in the Methods section.
A description of all covariates tested A description of any assumptions or corrections, such as tests of normality and adjustment for multiple comparisons A full description of the statistical parameters including central tendency (e.g. means) or other basic estimates (e.g. regression coefficient) AND variation (e.g. standard deviation) or associated estimates of uncertainty (e.g. confidence intervals) For null hypothesis testing, the test statistic (e.g. F, t, r) with confidence intervals, effect sizes, degrees of freedom and P value noted Give P values as exact values whenever suitable.

For Bayesian analysis, information on the choice of priors and Markov chain Monte Carlo settings
For hierarchical and complex designs, identification of the appropriate level for tests and full reporting of outcomes Estimates of effect sizes (e.g. Cohen's d, Pearson's r), indicating how they were calculated Our web collection on statistics for biologists contains articles on many of the points above.

Software and code
Policy information about availability of computer code Data collection

Data analysis
For manuscripts utilizing custom algorithms or software that are central to the research but not yet described in published literature, software must be made available to editors/reviewers. We strongly encourage code deposition in a community repository (e.g. GitHub). See the Nature Research guidelines for submitting code & software for further information.

Data
Policy information about availability of data All manuscripts must include a data availability statement. This statement should provide the following information, where applicable: -Accession codes, unique identifiers, or web links for publicly available datasets -A list of figures that have associated raw data -A description of any restrictions on data availability Field-specific reporting Please select the one below that is the best fit for your research. If you are not sure, read the appropriate sections before making your selection. All analysis code is custom written in Matlab and is freely available from the authors upon request.
The datasets analyzed during the current study are available in the Human Connectome Project database: https://www.humanconnectome.org, as well as genetic expression data freely available from the Allen Brain Institute.

nature research | reporting summary
October 2018 For a reference copy of the document with all sections, see nature.com/documents/nr-reporting-summary-flat.pdf Behavioural & social sciences study design All studies must disclose on these points even when the disclosure is negative. Note that full information on the approval of the study protocol must also be provided in the manuscript.

Magnetic resonance imaging Experimental design
Design type All analyses were quantitative measures of individual differences in human brain functional connectivity measured with fMRI, and individual differences in task performance on the n-back working memory task. Genetic data included genetic coexpression between a selection of genes across the brain.
All magnetic resonance imaging data analyzed is included in the Human Connectome Project S900 Release, all genetic expression data is available from the Allen Brain Institute No sample size calculation was employed.
All data was independently collected by the Human Connectome Project or the Allen Brain Institute. We analyzed all subjects that were part of the S900 release, as well as 6 postmortem samples Data was collected between 2009 and 2015.
We excluded all subjects in the S900 release that did not have all 4 of the following data: working memory task, resting state functional magnetic resonance scan, high resolution anatomical scan, diffusion tensor imaging scan. After excluding these subjects, our sample included 644 subjects.
For the fMRI data analyzed, only subjects that underwent (1) the working memory task, (2) resting state functional imaging, (3) high resolution anatomical imaging, and (4) diffusion tensor imaging were included. After excluding all subjects that did not complete each of these components, we were left with 644 subjects.
Subjects were not randomized to different groups. Anatomical location(s) The following descriptions for each task have been adapted for brevity from the Human Connectome Project Manual. Working Memory. The category specific representation task and the working memory task are combined into a single task paradigm. Participants were presented with blocks of trials that consisted of pictures of places, tools, faces and body parts (non-mutilated parts of bodies with no "nudity"). Within each run, the 4 different stimulus types were presented in separate blocks. Also, within each run, 1"2 of the blocks use a 2-back working memory task and 1"2 use a 0-back working memory task (as a working memory comparison). A 2.5 second cue indicates the task type (and target for 0-back) at the start of the block. Each of the two runs contains 8 task blocks (10 trials of 2.5 seconds each, for 25 seconds) and 4 fixation blocks (15 seconds). On each trial, the stimulus is presented for 2 seconds, followed by a 500 ms inter-task interval (ITI).
All performance measures were chosen a priori. In the working memory tasks, we used the mean accuracy across all nback conditions (face, body, place, tool), as well as the d-prime statistic.

3T
The acquisition parameters for each data type are as follows. The parameters for the acquisition of high-resolution structural scan were: TR = 2400 ms, TE = 2.14 ms, TI = 1000 ms, Registration of the T1 to atlas space includes an initial volumetric registration to MNI152 space using FSL's linear FLIRT tool, followed by the nonlinear FNIRT algorithm.

MNI152
For both resting-state and task functional connectivity, CompCor, with five principal components from the ventricles and white matter masks, was used to regress out nuisance signals from the time series. In addition, the 12 detrended motion estimates provided by the Human Connectome Project were regressed out from the time series. The mean global signal was removed and then time series were band-pass filtered from 0.009 to 0.08 Hz.
Frames with greater than 0.2 mm frame-wise displacement or a derivative root mean square (DVARS) above 75 were removed as outliers. Sessions composed of greater than 50 percent outlier frames were not further analyzed.

Pearson r correlations between all ROIs
We parcellated the brain into 400 discrete and non-overlapping regions of interest using the Schaefer atlas (fslr32k surface) [69]. Notably, the Schaefer atlas was originally validated in the same HCP data that we study here, and it yields a functional demarcation of both the default mode and the frontoparietal systems. Of course other functionally defined atlases exist, but they are less ideal for our purposes for several reasons; the Power atlas [64] does not provide full cortical coverage, and the Gordon [40] and Brainnetome [31] atlases are lower spatial resolution including 333 and 246 regions, respectively. The Schaefer atlas provides an assignment of each region to one of 17 putative cognitive systems: two visual, two somatomotor, two dorsal attention, two salience/ventral attention, one limbic, three frontoparietal, three default mode, and one temporo-parietal system. To ensure that the granularity of the data was consistent with the granularity of our hypotheses, we collapsed these 17 systems into 8 systems by combining individual systems that belonged to the same cognitive system; that is, we combined the two visual systems into a single system, the two somatomotor systems into a single system, the two dorsal attention systems into a single system, the two salience systems into a single system, the three