Dynamic fMRI networks of human emotion

The relationship between emotions and brain function has been a central focus of affective neuroscience, with foundational studies elucidating the roles of structures such as the amygdala, insula, and prefrontal cortex in emotional processing (LaBar et al., 1995; Ochsner et al., 2002). Despite significant progress, most functional MRI (fMRI) studies have adopted a static view of emotions, often emphasizing localized brain activations or static connectivity patterns rather than the dynamic and distributed nature of emotional processing (see Kuppens and Verduyn, 2017; Pessoa, 2017; Waugh et al., 2015, for discussion). This particular approach to studying emotions arises in part because analytical techniques collapse time-varying fMRI activity into static spatial maps and also because experimental paradigms typically use unchanging stimuli (Murphy et al., 2003; Phan et al., 2002). In the present study, we collected fMRI data while participants performed a behavioral task with trials that paired the viewing of emotionally evocative short movie clips with a subsequent discrete emotion decision task, as well as an intertrial rest period. We sought to address three key questions: (1) What distinct whole-brain networks underlie performance during this task? (2) Do the spatial distributions of these networks vary as a function of the emotional content of the stimuli? (3) Are the temporal dynamics of these networks—such as their time to peak, peak value, or duration—sensitive to the emotional characteristics of the stimuli? To answer these questions, we relied on a data-driven approach that emphasized a combined spatial and temporal view on fMRI data analysis.

Previous studies investigating emotional processing with fMRI have generally revealed a static view of emotions in the brain. Specifically, recent research has highlighted that emotions engage widespread networks involving multiple interconnected regions, reflecting the complex and distributed nature of emotional processing (Kober et al., 2008; Lindquist et al., 2012; Palomero-Gallagher and Amunts, 2022; Pessoa, 2017; Zhou et al., 2021). For instance, Riedel et al., 2018, applied a meta-analytic clustering approach to over 1700 neuroimaging experiments to delineate five distinct whole-brain networks associated with emotional processing. The first two networks were linked with visual and auditory perception, with convergent activation in the visual (occipital) and auditory (superior temporal) cortices, respectively. The third network captured the salience network, including the insula and dorsal anterior cingulate cortex, reflecting attention to emotionally salient information. The fourth network was associated with appraisal and prediction of emotional events, involving regions of the default mode network, such as the medial prefrontal and posterior cingulate cortices. The fifth network focused on the induction of emotional responses, with key contributions from the amygdala, parahippocampal gyri, and fusiform gyri, underscoring its role in generating contextually relevant emotional responses. While this and other studies highlight the involvement of large-scale networks in the processing of emotion and therefore go beyond the traditional one-to-one mapping between brain regions and emotions (e.g. Murphy et al., 2003), they do not reveal the temporal component associated with these networks.

Traditionally, many fMRI studies of emotion have employed decontextualized static images of facial expressions or evocative pictures to elicit emotional responses (Adolphs, 2002; Liu et al., 2021; Wallenwein et al., 2024). In contrast, more recent studies have argued that movies offer a richer perceptual experience that is of a higher ecological validity (Sonkusare et al., 2019; Finn and Bandettini, 2021). Filmmakers are adept at eliciting specific emotions in viewers (Westermann et al., 1996), and recent studies have shown that using movies as emotion-evoking stimuli can provide new insights into the relationship between emotions and the brain (Jääskeläinen et al., 2021; Morgenroth et al., 2023; Saarimäki, 2021). However, a limitation of current studies using movie clips is that they often employ lengthy segments lasting several minutes, during which multiple cognitive and emotional events can occur (e.g. Xu et al., 2023; Vemuri and Surampudi, 2015). This complexity makes it challenging to disentangle the dynamic changes in emotional processing over time (but see Meer et al., 2020). To address this issue, shorter movie clips focusing on a single event can be utilized to better control and observe how emotions develop over time within events. In addition, an important aspect of emotions concerns not only their initial expression but also how they are regulated following that initial response (Etkin et al., 2015; Ochsner and Gross, 2008). This suggests that the study of the brain dynamics of emotions should focus not only on the stimulus event but should also examine how brain responses change after the event has concluded. Incorporating short periods of rest after the presentation of short movie clips may facilitate the investigation of post-event emotional regulation processes and how they are represented in the brain.

In the current study, participants viewed 12.5 s audiovisual clips selected from the movie Forrest Gump (Zemeckis et al., 1994). Each clip was specifically chosen to evoke a single emotion—happiness, fear, or sadness—with each emotion equally represented across a total of 15 clips. Following each clip, participants completed a two-alternative forced-choice decision task using discrete emotion words denoting the emotion experienced by the movie’s protagonist. Finally, there was a 10 s intertrial rest period. We collected whole-brain fMRI data at 2 mm spatial resolution, and extracted the event-related fMRI signal at 2 s temporal resolution across a 28 s epoch for each emotion type and for each participant using slice-based fMRI (Janssen et al., 2018; Janssen and Mendieta, 2020). In parallel with resting-state analyses, we then performed group spatial independent component analysis (ICA) to identify large-scale networks with distinct spatial distributions and temporal profiles (Beckmann and Smith, 2004; Smith et al., 2009). The spatial and temporal properties of the discovered networks were further quantified using dual regression (Beckmann et al., 2009) and Gaussian curve fitting (Janssen and Mendieta, 2020; Kruggel et al., 2000; Lindquist et al., 2009), respectively. Finally, we performed mixed-effects regression modeling and inferential statistics to establish the reliability of our results.

The emotions that are experienced by the protagonist in each clip developed gradually over the course of each clip rather than appearing immediately. In other words, while the visual and auditory perception of the movie clip is instantaneous, the comprehension of its emotional content requires more nuanced processing that lags behind the immediate perceptual input stream. Therefore, we expected to detect brain networks associated with initial sensory processing, involving primary visual and auditory cortices (Riedel et al., 2018), as well as networks related to higher-level cognitive and emotional processing, such as those involving the middle temporal gyrus and other association areas (Huth et al., 2016; Visser et al., 2012). Given that we analyzed the entire behavioral task—including the decision stage and the rest period—we anticipated identifying additional large-scale networks associated with these later stages of the task, such as a response initiation network involving primary motor cortices and the default mode network (Satpute and Lindquist, 2019).

Analysis of the behavioral data showed that participants did not treat the three emotion categories in the same way. Specifically, as can be seen in Figure 1C, there was higher agreement among participants on the target response in the sad vs the other two emotion categories (Sad vs Happy p<0.02; Sad vs Fear p<0.04). In addition, as can be seen in Figure 1D, decisions for the sad condition were faster compared to the other two conditions (Sad vs Happy p<0.003; Sad vs Fear p<0.008; Figure 2).

Figure 1

Download asset Open asset

Figure 2

Download asset Open asset

Group spatial ICA of the 4D slice-based fMRI datasets first revealed that the estimated optimal model selection occurred with six dimensions. As can be seen in Figure 3, four of these networks corresponded to signal, and two were considered noise. Figure 3 shows both spatial distributions and mean temporal profiles associated with each IC network. Visual inspection of these temporal profiles enables an association with the different stages of the behavioral task. Specifically, IC0 and IC1 show an activation pattern consistent with the movie watching task where there was an immediate onset, a sustained activity, and offset consistent with the presentation of the 12.5 s movie clips. Similarly, IC2 shows a pattern associated with the emotion decision stage with a peak in the decision stage of the task. In addition, the spatial distribution of co-activities of IC4 is consistent with the default mode network (dmn). For expository reasons, we labeled the IC networks in terms of a tentative functional interpretation. Specifically, for reasons we discuss in more detail below, we labeled IC1 with ‘input’, IC0 with ‘meaning’, IC2 with ‘response’, and IC4 with ‘dmn’. We also ordered these IC networks on the basis of their assumed temporal order of activation (i.e. IC1→IC0→IC4→IC2). Finally, IC3 and IC5 were classified as noise due to the presence of ‘activation’ in non-gray-matter areas of the brain (CSF, veins; Griffanti et al., 2014; Table 1).

Figure 3

Download asset Open asset

Further investigation of the spatial maps using dual regression and statistical modeling revealed insight into the specific regions that were functionally co-activated within each IC network. Specifically, modeling of the functional connectivity data with Equation 1 revealed a significant interaction between $IC\mathrm{\_}network$ and Region (p<0.0001; see Table 2 for all stats), suggesting that the pattern of regional functional connectivity was not the same across the four IC networks. Post hoc analyses with Bonferroni correction revealed patterns of regional functional connectivity that were unique to each IC network. These patterns are visualized in Figures 4 and 5. Whereas both IC0 (‘meaning’) and IC1 (‘input’) strongly relied on visual areas and temporal areas (lateral occipital, pericalcarine gyrus), IC1 (‘input’) differed from IC0 (‘meaning’) in its co-activity in the superior temporal gyrus, amygdala, and medial posterior cingulate regions. The IC2 (‘response’) network had strong activity in the primary motor regions as well as the anterior cingulate. Finally, the IC4 (‘dmn’) network showed strong activity in typical default mode network regions.

Figure 4

Download asset Open asset

Figure 5

Download asset Open asset

Statistical modeling with Equation 1 also revealed a triple interaction between $IC\mathrm{\_}network$, Region, and $Emotion\mathrm{\_}Condition$ (p<0.0001; see Table 2 for all stats), suggesting that the interaction between $IC\mathrm{\_}network$ and Region outlined above was further modulated by the Emotion Condition. Post hoc analyses of this three-way interaction revealed different effects of Emotion Condition for different regions within each IC network. As can be seen in Figure 6, discrete emotions modulated regions across the whole brain within all four IC networks. Specifically, discrete emotions modulated functional connectivity in visual, temporal, frontal, parietal, and subcortical regions for all four networks.

Figure 6

Download asset Open asset

Curve fitting of a Gaussian function to the temporal profiles associated with the four IC networks revealed further insight into the temporal properties of the four networks. First, the results revealed relatively high R2 values, suggesting that goodness of fit was adequate across the four IC networks (see Figure 7B). In addition, statistical modeling with Equation 3 revealed a main effect of $IC\mathrm{\_}network$ for all three dependent variables, suggesting that peak values, time to peak, and durations were not equal for the four IC networks (see Table 3 for details). Post hoc analyses of $IC\mathrm{\_}network$ for each dependent variable revealed that peak values differed between the four networks, where time-to-peak values were earlier for IC1 (‘input’) than for IC0 (‘meaning’), and durations were generally shorter for the networks associated in the later aspects of the task (see Figure 7C, D, and E for further details).

Figure 7

Download asset Open asset

Finally, statistical modeling with Equation 3 also revealed an interaction between $IC\mathrm{\_}network$ and $Emotion\mathrm{\_}Condition$ for all three dependent variables, suggesting that peak value, time to peak, and duration values for the happy, sad, and fear conditions were not equal across the four IC networks (see Table 3 for details). Post hoc analyses revealed that the Emotion Condition modulated peak value and duration for IC1 (‘input’), peak value and time to peak for IC0 (‘meaning’), time to peak for IC2 (‘response’), and duration for IC4 (‘dmn’; see Figure 8A–C for additional details).

Figure 8

Download asset Open asset

In the current study, we relied on time-resolved fMRI to discover the large-scale networks associated with a behavioral task that consisted of three stages: watching a movie clip (12.5 s), making a decision about the emotional experience by the protagonist in the clip (2.5 s), and an intertrial rest period (10 s). Group spatial ICA of slice-based fMRI data with increased temporal resolution across the entire trial period (28 s) revealed four large-scale brain networks that were associated with different aspects of the task. Two networks (IC0 and IC1) were associated with the movie stimulus, while a third network (IC2) activated during the later decision stage, and a fourth network (IC4) was activated throughout the task and showed a peak during the intertrial rest period. Further quantification of the spatial and temporal properties of these networks using dual regression and Gaussian curve fitting showed that the emotional content of the clip modulated the spatial distribution, as well as the peak value, time to peak, and duration properties of each network.

Brain networks IC0 and IC1 appear to play complementary roles during the movie-watching phase. IC1, which we labeled the ‘input’ network, aligns with aspects of Networks 1 and 2 of Riedel et al., 2018, involving primary visual and auditory processing regions critical for perceiving the movie stimulus (see Figure 4). However, IC1 (‘input’) also showed a resemblance to Network 5 of Riedel et al., 2018, with activity in the amygdala, fusiform gyrus, and parahippocampal gyrus that was linked to the evaluation of the emotional content of the stimulus. Note that no other network detected in our data consisted of co-activity in these three regions. In addition, time-to-peak values were earlier in the IC1 (‘input’) network than in all other networks (see Figure 7D). The combined spatial and temporal properties of this network therefore suggest that this network is likely responsible for processing the visual and auditory input stream, as well as the evaluation of the emotional content of the stimulus.

On the other hand, although also associated with the movie stimulus, the IC0 (‘meaning’) network appeared to have spatial and temporal properties that were distinct from the IC1 (‘input’) network. Specifically, IC0 (‘meaning’) had increased co-activity in lateral frontal and middle temporal areas that are typically associated with semantic processing (see Figure 5; Huth et al., 2016; Visser et al., 2012). In addition, IC0 (‘meaning’) aligned with Network 3 of Riedel et al., 2018, and appeared to contain regions associated with the salience network such as the insula and the anterior cingulate gyrus (Seeley, 2019), suggesting an attentional component. Importantly, even though the IC0 (‘meaning’) network activated in response to the movie stimulus, its time-to-peak values lagged about 2 s behind those of IC1 (‘input’; see Figure 7). As was pointed out in the Introduction, the movie clips showed scenes whose intended meaning could not be discerned immediately, but instead required additional processing that lagged behind the input stream. As before, the combined spatial and temporal characteristics of this network allow us to interpret this network as involved in the processing of the meaning associated with the input stream.

In addition, a third network (IC2, ‘response’) showed temporal activity with a peak around 20 s and can therefore be associated with the response stage (see Figure 3). The spatial properties of this network showed activity in cortical and subcortical regions associated with processing the visual input (the discrete emotion words), decision-making-associated regions such as the caudate nucleus (Grahn et al., 2008), motor generation regions such as the precentral gyrus, and task general attention-related regions such as the insula and the anterior cingulate cortex (Seeley, 2019). This therefore suggests that this network represents the combined emotion decision and motor output network. Note that with increased temporal resolution (and potentially restricting behavioral variance in response latency), one might be able to separate the decision and the motor output components of this network. Finally, the fourth network (IC4, ‘dmn’) showed an alignment with Network 4 of Riedel et al., 2018, and had spatial properties consistent with those of the default mode network (Buckner et al., 2008). Although the exact role of the default mode network in emotion processing is currently debated (Satpute and Lindquist, 2019), it is noteworthy that the network showed some co-activity throughout the task and had a peak during the rest period (see Figure 3), and showed activity in lateral prefrontal regions (see Figure 4). These spatial and temporal properties are consistent with an interpretation in terms of the regulation, mind-wandering, and/or self-referential processing of the emotions (Gross, 2015).

Both the spatial and temporal properties of the four identified networks were influenced by the emotional content of the movie clips. Specifically, in IC1 (‘input’), regions in the visual, parietal, and frontal lobes showed differential activation patterns across happy, fear, and sad clips. Temporal dynamics, including the duration and peak values of the responses, also varied between these emotional conditions (Figures 6A and 8A). Similar modulations were observed for IC0 (‘meaning’), IC2 (‘response’), and IC4 (‘dmn’), where emotional content influenced activations in visual, temporal, and frontal regions, as well as temporal metrics such as response duration, time to peak, and peak amplitude. Notably, the time-to-peak differences in IC2 (‘response’) align closely with response latency results, with sad trials showing faster response latencies and earlier peak times (compare Figures 1D and 8D). Critically, the same ordering between response latencies and the fMRI time-to-peak metric provides convergent temporal validity for using BOLD dynamics as a window into the speed of affective computations. Nevertheless, these results should not be taken as evidence that processing sad emotions is generally faster than other emotions. Other studies have demonstrated fast processing of fearful stimuli (Méndez-Bértolo et al., 2016; Bo et al., 2022; Grootswagers et al., 2020), and specific aspects of the current experimental paradigm likely influenced the speed of emotional processing. In general, the current findings suggest that emotional stimuli elicit widespread effects on brain activity, impacting not only the processing of emotion-laden input stimuli but also the spatial and temporal dynamics of broader neural responses.

Taken together, the results of this study paint a comprehensive picture of how the brain orchestrates spatially and temporally distinct networks to process emotional stimuli and execute task-relevant behaviors. The findings highlight a dynamic interplay between sensory processing, semantic interpretation, decision-making, and emotional regulation, all of which unfold within a distributed but highly coordinated network architecture. Although the networks demonstrated separable time courses, their temporal dynamics were largely overlapping, suggesting that processes such as those mediated by IC1 (‘input’) and IC0 (‘meaning’) likely operate in parallel rather than in isolation. This parallel engagement implies that cross-talk between networks is a critical feature of task-related brain activity (Cole et al., 2013), particularly during complex cognitive and emotional tasks. For example, the interplay between sensory-driven processing in IC1 and the semantic and attentional mechanisms in IC0 may be necessary for the integrated perception of emotionally salient stimuli. How such cross-talk is mediated remains an open question, representing an intriguing avenue for future research. In the decision phase, IC2 (‘response’) reveals how emotion processing transitions from perception to cognitive evaluation, involving a convergence of sensory, attentional, and motor networks to guide task execution. Finally, IC4, (‘dmn’) associated with the default mode network, demonstrates the presence of reflective, mind-wandering, and/or regulatory processes (Gross, 2015), particularly during rest but with influence throughout the task (see Figure 3). These findings underscore the dynamic nature of brain activity, where emotional stimuli engage overlapping networks that continuously integrate sensory input, semantic processing, decision-making, and regulation. This fluid coordination reveals how emotions are adaptively represented and processed across time and space (Waugh et al., 2015; Pessoa and McMenamin, 2017; John et al., 2022).

Our study has several limitations that warrant discussion. First, the interpretation of BOLD signal dynamics is inherently influenced by the vascular properties of the underlying tissue (Ogawa et al., 1992). This limitation has been discussed primarily in the context of studies that have attempted to determine the temporal sequence of activation of different brain regions where such differences could be confounded by variations in vascular responsiveness (Aguirre et al., 1998; Kim et al., 1997; Menon et al., 1998). However, it is important to note that the temporal properties of the signal we analyzed do not represent those of individual brain regions but the aggregate activity of many regions working in concert, reducing the likelihood of significant confounding effects. A further limitation is that we did not require participants to directly regulate their emotions, which complicates an interpretation in terms of emotion regulation (Gross, 2015). Likewise, the current research design only used three main emotional categories. Whether our conclusions hold for other emotions and emotion types remains to be studied in future studies. Another general limitation of the results presented here concerns the influence of the number of dimensions chosen in group spatial ICA on the spatial distribution of the observed large-scale networks. It is well documented that increasing the number of dimensions can lead to the ‘splitting’ of networks into finer components, while fewer dimensions may overgeneralize distinct functional patterns (Beckmann, 2012). This presents a challenge, as there is no a priori standard for the ‘correct’ spatial distribution of task-related networks, especially for complex tasks such as watching movie clips (Di et al., 2013; Dosenbach et al., 2025; Huang et al., 2024). In this sense, task-based fMRI studies of large-scale networks are currently in a similar position to early resting-state fMRI research, where the precise spatial distributions of networks were still being defined (e.g. Yeo et al., 2011). To address this, a concerted effort is needed to map the large-scale networks underlying diverse tasks, enabling more informed choices about dimensionality and mitigating the uncertainty in network characterization.

Finally, it is worth emphasizing that our approach bears a relationship to what is often termed ‘dynamic functional connectivity’ (Gonzalez-Castillo and Bandettini, 2018; Hutchison et al., 2013). In particular, our results demonstrate how whole-brain functional connectivity maps shift over time in response to different components of a complex, emotion-based behavioral task (Figure 3). Although both event-related fMRI and group spatial ICA have been employed extensively in prior studies, combining them as in the current study is less common (Janssen and Mendieta, 2020; Long et al., 2013). The particular slice-based event-related approach has been shown to offer enhanced temporal precision and greater flexibility in choosing different temporal resolutions than more traditional signal extraction schemes (Janssen et al., 2018). Indeed, when we analyzed our data at lower temporal resolution (i.e. at the 3 s TR level), some components with insufficient temporal separation (such as IC1 and IC0) became conflated, highlighting the importance of higher-resolution signal extraction in distinguishing these networks (Calhoun et al., 2001). Moreover, we used standard group spatial ICA methods, which have been widely applied to resting-state data, to derive task-based whole-brain functional connectivity maps. This unified approach for finding networks in task-based and resting-state data may facilitate comparisons between networks obtained from different modalities (Gonzalez-Castillo and Bandettini, 2018). Taken together, these findings demonstrate the value of combining slice-based event-related fMRI with group spatial ICA to characterize how functional connectivity patterns evolve over time. Future research may further explore the interplay of the temporal resolution of the extracted signal, group spatial ICA, and alternative approaches, such as FIR analysis (Josephs et al., 1997; Windischberger et al., 2008), to illuminate the spatiotemporal characteristics of dynamic network changes during complex tasks.

To conclude, the present study addressed three key questions: whether distinct whole-brain networks underlie the different stages of our emotional movie watching and decision task, whether their spatial distributions vary with emotional content, and whether their temporal dynamics—including time-to-peak and duration—are sensitive to emotion. Our results revealed four large-scale networks whose activity patterns not only differentiated task stages but revealed a separation between a network related to processing affective input and understanding the meaning of this input. Importantly, the spatial and temporal features of all four networks varied as a function of the emotional content of the stimuli. These results move beyond static, region-based models and suggest emotion reflects dynamic, distributed interactions among multiple networks (Lindquist et al., 2012; Pessoa, 2017). In turn, our findings underscore the utility of examining temporal metrics to capture subtle nuances of emotional processing that may remain undetectable using standard static analyses. These temporal properties may be especially relevant for clinical studies in which certain emotional disorders arise due to disturbances in the temporal regulation of emotion (e.g. dwelling too long on certain thoughts; Kuppens and Verduyn, 2017; Waugh et al., 2015). The current study’s approach for quantifying these temporal properties can therefore offer valuable insights, both for basic science and for the development of novel clinical tools.

The data and analysis code supporting the findings of this study are publicly available on the Open Science Framework at https://doi.org/10.17605/OSF.IO/XY43A.

1. 1. Janssen N

   (2025) Open Science Framework

   Dynamic fMRI networks of emotion.

   https://doi.org/10.17605/OSF.IO/XY43A

1. 1. Adolphs R

   (2002) Recognizing emotion from facial expressions: psychological and neurological mechanisms

   Behavioral and Cognitive Neuroscience Reviews 1:21–62.

   https://doi.org/10.1177/1534582302001001003

   • PubMed
   • Google Scholar
2. 1. Aguirre GK
   2. Zarahn E
   3. D’esposito M

   (1998) The variability of human, BOLD hemodynamic responses

   NeuroImage 8:360–369.

   https://doi.org/10.1006/nimg.1998.0369

   • PubMed
   • Google Scholar
3. 1. Bates D
   2. Mächler M
   3. Bolker B
   4. Walker S

   (2015) Fitting linear mixed-effects models using lme4

   Journal of Statistical Software 67:1–48.

   https://doi.org/10.18637/jss.v067.i01

   • Google Scholar
4. 1. Beckmann CF
   2. Smith SM

   (2004) Probabilistic independent component analysis for functional magnetic resonance imaging

   IEEE Transactions on Medical Imaging 23:137–152.

   https://doi.org/10.1109/TMI.2003.822821

   • PubMed
   • Google Scholar
5. 1. Beckmann CF
   2. Mackay CE
   3. Filippini N
   4. Smith SM

   (2009) Group comparison of resting-state FMRI data using multi-subject ICA and dual regression

   NeuroImage 47:S148.

   https://doi.org/10.1016/S1053-8119(09)71511-3

   • Google Scholar
6. 1. Beckmann CF

   (2012) Modelling with independent components

   NeuroImage 62:891–901.

   https://doi.org/10.1016/j.neuroimage.2012.02.020

   • PubMed
   • Google Scholar
7. 1. Bo K
   2. Cui L
   3. Yin S
   4. Hu Z
   5. Hong X
   6. Kim S
   7. Keil A
   8. Ding M

   (2022) Decoding the temporal dynamics of affective scene processing

   NeuroImage 261:119532.

   https://doi.org/10.1016/j.neuroimage.2022.119532

   • PubMed
   • Google Scholar
8. 1. Buckner RL
   2. Andrews-Hanna JR
   3. Schacter DL

   (2008) The brain’s default network: anatomy, function, and relevance to disease

   Annals of the New York Academy of Sciences 1124:1–38.

   https://doi.org/10.1196/annals.1440.011

   • PubMed
   • Google Scholar
9. 1. Calhoun VD
   2. Adali T
   3. Pearlson GD
   4. Pekar JJ

   (2001) Spatial and temporal independent component analysis of functional MRI data containing a pair of task-related waveforms

   Human Brain Mapping 13:43–53.

   https://doi.org/10.1002/hbm.1024

   • PubMed
   • Google Scholar
10. 1. Cole MW
    2. Reynolds JR
    3. Power JD
    4. Repovs G
    5. Anticevic A
    6. Braver TS

    (2013) Multi-task connectivity reveals flexible hubs for adaptive task control

    Nature Neuroscience 16:1348–1355.

    https://doi.org/10.1038/nn.3470

    • PubMed
    • Google Scholar
11. 1. Desikan RS
    2. Ségonne F
    3. Fischl B
    4. Quinn BT
    5. Dickerson BC
    6. Blacker D
    7. Buckner RL
    8. Dale AM
    9. Maguire RP
    10. Hyman BT
    11. Albert MS
    12. Killiany RJ

    (2006) An automated labeling system for subdividing the human cerebral cortex on MRI scans into gyral based regions of interest

    NeuroImage 31:968–980.

    https://doi.org/10.1016/j.neuroimage.2006.01.021

    • PubMed
    • Google Scholar
12. 1. Di X
    2. Gohel S
    3. Kim EH
    4. Biswal BB

    (2013) Task vs. rest-different network configurations between the coactivation and the resting-state brain networks

    Frontiers in Human Neuroscience 7:493.

    https://doi.org/10.3389/fnhum.2013.00493

    • PubMed
    • Google Scholar
13. 1. Dosenbach NUF
    2. Raichle ME
    3. Gordon EM

    (2025) The brain’s action-mode network

    Nature Reviews. Neuroscience 26:158–168.

    https://doi.org/10.1038/s41583-024-00895-x

    • PubMed
    • Google Scholar
14. 1. Duann JR
    2. Jung TP
    3. Kuo WJ
    4. Yeh TC
    5. Makeig S
    6. Hsieh JC
    7. Sejnowski TJ

    (2002) Single-trial variability in event-related BOLD signals

    NeuroImage 15:823–835.

    https://doi.org/10.1006/nimg.2001.1049

    • PubMed
    • Google Scholar
15. Software

    1. Elzhov TV
    2. Mullen KM
    3. Spiess AN
    4. Bolker B

    (2022) Minpack.lm: r interface to the levenberg-marquardt nonlinear least-squares algorithm, version Version 1.2-4

    R Package.

    https://doi.org/10.32614/CRAN.package.minpack.lm
16. 1. Etkin A
    2. Büchel C
    3. Gross JJ

    (2015) The neural bases of emotion regulation

    Nature Reviews. Neuroscience 16:693–700.

    https://doi.org/10.1038/nrn4044

    • PubMed
    • Google Scholar
17. 1. Finn ES
    2. Bandettini PA

    (2021) Movie-watching outperforms rest for functional connectivity-based prediction of behavior

    NeuroImage 235:117963.

    https://doi.org/10.1016/j.neuroimage.2021.117963

    • PubMed
    • Google Scholar
18. 1. Fischl B

    (2012) FreeSurfer

    NeuroImage 62:774–781.

    https://doi.org/10.1016/j.neuroimage.2012.01.021

    • PubMed
    • Google Scholar
19. 1. Glasser MF
    2. Sotiropoulos SN
    3. Wilson JA
    4. Coalson TS
    5. Fischl B
    6. Andersson JL
    7. Xu J
    8. Jbabdi S
    9. Webster M
    10. Polimeni JR
    11. Van Essen DC
    12. Jenkinson M
    13. WU-Minn HCP Consortium

    (2013) The minimal preprocessing pipelines for the Human Connectome Project

    NeuroImage 80:105–124.

    https://doi.org/10.1016/j.neuroimage.2013.04.127

    • PubMed
    • Google Scholar
20. 1. Gonzalez-Castillo J
    2. Bandettini PA

    (2018) Task-based dynamic functional connectivity: recent findings and open questions

    NeuroImage 180:526–533.

    https://doi.org/10.1016/j.neuroimage.2017.08.006

    • PubMed
    • Google Scholar
21. 1. Grahn JA
    2. Parkinson JA
    3. Owen AM

    (2008) The cognitive functions of the caudate nucleus

    Progress in Neurobiology 86:141–155.

    https://doi.org/10.1016/j.pneurobio.2008.09.004

    • PubMed
    • Google Scholar
22. 1. Griffanti L
    2. Salimi-Khorshidi G
    3. Beckmann CF
    4. Auerbach EJ
    5. Douaud G
    6. Sexton CE
    7. Zsoldos E
    8. Ebmeier KP
    9. Filippini N
    10. Mackay CE
    11. Moeller S
    12. Xu J
    13. Yacoub E
    14. Baselli G
    15. Ugurbil K
    16. Miller KL
    17. Smith SM

    (2014) ICA-based artefact removal and accelerated fMRI acquisition for improved resting state network imaging

    NeuroImage 95:232–247.

    https://doi.org/10.1016/j.neuroimage.2014.03.034

    • PubMed
    • Google Scholar
23. 1. Grootswagers T
    2. Kennedy BL
    3. Most SB
    4. Carlson TA

    (2020) Neural signatures of dynamic emotion constructs in the human brain

    Neuropsychologia 145:106535.

    https://doi.org/10.1016/j.neuropsychologia.2017.10.016

    • PubMed
    • Google Scholar
24. 1. Gross JJ

    (2015) The extended process model of emotion regulation: elaborations, applications, and future directions

    Psychological Inquiry 26:130–137.

    https://doi.org/10.1080/1047840X.2015.989751

    • Google Scholar
25. 1. Huang S
    2. De Brigard F
    3. Cabeza R
    4. Davis SW

    (2024) Connectivity analyses for task-based fMRI

    Physics of Life Reviews 49:139–156.

    https://doi.org/10.1016/j.plrev.2024.04.012

    • PubMed
    • Google Scholar
26. 1. Hutchison RM
    2. Womelsdorf T
    3. Allen EA
    4. Bandettini PA
    5. Calhoun VD
    6. Corbetta M
    7. Della Penna S
    8. Duyn JH
    9. Glover GH
    10. Gonzalez-Castillo J
    11. Handwerker DA
    12. Keilholz S
    13. Kiviniemi V
    14. Leopold DA
    15. de Pasquale F
    16. Sporns O
    17. Walter M
    18. Chang C

    (2013) Dynamic functional connectivity: promise, issues, and interpretations

    NeuroImage 80:360–378.

    https://doi.org/10.1016/j.neuroimage.2013.05.079

    • PubMed
    • Google Scholar
27. 1. Huth AG
    2. de Heer WA
    3. Griffiths TL
    4. Theunissen FE
    5. Gallant JL

    (2016) Natural speech reveals the semantic maps that tile human cerebral cortex

    Nature 532:453–458.

    https://doi.org/10.1038/nature17637

    • PubMed
    • Google Scholar
28. 1. Jääskeläinen IP
    2. Sams M
    3. Glerean E
    4. Ahveninen J

    (2021) Movies and narratives as naturalistic stimuli in neuroimaging

    NeuroImage 224:117445.

    https://doi.org/10.1016/j.neuroimage.2020.117445

    • PubMed
    • Google Scholar
29. 1. Janssen N
    2. Hernández-Cabrera JA
    3. Foronda LE

    (2018) Improving the signal detection accuracy of functional Magnetic Resonance Imaging

    NeuroImage 176:92–109.

    https://doi.org/10.1016/j.neuroimage.2018.01.076

    • PubMed
    • Google Scholar
30. 1. Janssen N
    2. Mendieta CCR

    (2020) The dynamics of speech motor control revealed with time-resolved fMRI

    Cerebral Cortex 30:241–255.

    https://doi.org/10.1093/cercor/bhz084

    • PubMed
    • Google Scholar
31. Software

    1. Janssen N

    (2022) Display_melodics, version a349c4a

    Github.

    https://github.com/iamnielsjanssen/display_melodics
32. 1. Jenkinson M
    2. Bannister P
    3. Brady M
    4. Smith S

    (2002) Improved optimization for the robust and accurate linear registration and motion correction of brain images

    NeuroImage 17:825–841.

    https://doi.org/10.1016/s1053-8119(02)91132-8

    • PubMed
    • Google Scholar
33. 1. John YJ
    2. Sawyer KS
    3. Srinivasan K
    4. Müller EJ
    5. Munn BR
    6. Shine JM

    (2022) It’s about time: linking dynamical systems with human neuroimaging to understand the brain

    Network Neuroscience 6:960–979.

    https://doi.org/10.1162/netn_a_00230

    • PubMed
    • Google Scholar
34. 1. Josephs O
    2. Turner R
    3. Friston K

    (1997) Event-related f MRI

    Human Brain Mapping 5:243–248.

    https://doi.org/10.1002/(SICI)1097-0193(1997)5:4<243::AID-HBM7>3.0.CO;2-3

    • PubMed
    • Google Scholar
35. 1. Kim SG
    2. Richter W
    3. Uğurbil K

    (1997) Limitations of temporal resolution in functional MRI

    Magnetic Resonance in Medicine 37:631–636.

    https://doi.org/10.1002/mrm.1910370427

    • PubMed
    • Google Scholar
36. 1. Kober H
    2. Barrett LF
    3. Joseph J
    4. Bliss-Moreau E
    5. Lindquist K
    6. Wager TD

    (2008) Functional grouping and cortical-subcortical interactions in emotion: a meta-analysis of neuroimaging studies

    NeuroImage 42:998–1031.

    https://doi.org/10.1016/j.neuroimage.2008.03.059

    • PubMed
    • Google Scholar
37. 1. Kruggel F
    2. von Cramon DY

    (1999) Modeling the hemodynamic response in single-trial functional MRI experiments

    Magnetic Resonance in Medicine 42:787–797.

    https://doi.org/10.1002/(sici)1522-2594(199910)42:4<787::aid-mrm22>3.0.co;2-v

    • PubMed
    • Google Scholar
38. 1. Kruggel F
    2. Zysset S
    3. von Cramon DY

    (2000) Nonlinear regression of functional MRI data: an item recognition task study

    NeuroImage 12:173–183.

    https://doi.org/10.1006/nimg.2000.0604

    • PubMed
    • Google Scholar
39. 1. Kuppens P
    2. Verduyn P

    (2017) Emotion dynamics

    Current Opinion in Psychology 17:22–26.

    https://doi.org/10.1016/j.copsyc.2017.06.004

    • PubMed
    • Google Scholar
40. 1. Kuznetsova A
    2. Brockhoff PB
    3. Christensen RHB

    (2017) lmerTest package: tests in linear mixed effects models

    Journal of Statistical Software 82:1–26.

    https://doi.org/10.18637/jss.v082.i13

    • Google Scholar
41. 1. LaBar KS
    2. LeDoux JE
    3. Spencer DD
    4. Phelps EA

    (1995) Impaired fear conditioning following unilateral temporal lobectomy in humans

    The Journal of Neuroscience 15:6846–6855.

    https://doi.org/10.1523/JNEUROSCI.15-10-06846.1995

    • PubMed
    • Google Scholar
42. Software

    1. Lenth R
    2. Singmann H
    3. Love J
    4. Buerkner P
    5. Herve M

    (2018) Emmeans: estimated marginal means, aka least-squares means, version Version 2.0.4

    R Package.

    https://doi.org/10.32614/CRAN.package.emmeans
43. 1. Lindquist MA
    2. Meng Loh J
    3. Atlas LY
    4. Wager TD

    (2009) Modeling the hemodynamic response function in fMRI: efficiency, bias and mis-modeling

    NeuroImage 45:S187–S198.

    https://doi.org/10.1016/j.neuroimage.2008.10.065

    • PubMed
    • Google Scholar
44. 1. Lindquist KA
    2. Wager TD
    3. Kober H
    4. Bliss-Moreau E
    5. Barrett LF

    (2012) The brain basis of emotion: a meta-analytic review

    The Behavioral and Brain Sciences 35:121–143.

    https://doi.org/10.1017/S0140525X11000446

    • PubMed
    • Google Scholar
45. 1. Liu M
    2. Liu CH
    3. Zheng S
    4. Zhao K
    5. Fu X

    (2021) Reexamining the neural network involved in perception of facial expression: A meta-analysis

    Neuroscience & Biobehavioral Reviews 131:179–191.

    https://doi.org/10.1016/j.neubiorev.2021.09.024

    • Google Scholar
46. 1. Long Z
    2. Li R
    3. Wen X
    4. Jin Z
    5. Chen K
    6. Yao L

    (2013) Separating 4D multi-task fMRI data of multiple subjects by independent component analysis with projection

    Magnetic Resonance Imaging 31:60–74.

    https://doi.org/10.1016/j.mri.2012.06.034

    • PubMed
    • Google Scholar
47. 1. Meer JNVD
    2. Breakspear M
    3. Chang LJ
    4. Sonkusare S
    5. Cocchi L

    (2020) Movie viewing elicits rich and reliable brain state dynamics

    Nature Communications 11:5004.

    https://doi.org/10.1038/s41467-020-18717-w

    • Google Scholar
48. 1. Méndez-Bértolo C
    2. Moratti S
    3. Toledano R
    4. Lopez-Sosa F
    5. Martínez-Alvarez R
    6. Mah YH
    7. Vuilleumier P
    8. Gil-Nagel A
    9. Strange BA

    (2016) A fast pathway for fear in human amygdala

    Nature Neuroscience 19:1041–1049.

    https://doi.org/10.1038/nn.4324

    • PubMed
    • Google Scholar
49. 1. Menon RS
    2. Luknowsky DC
    3. Gati JS

    (1998) Mental chronometry using latency-resolved functional MRI

    PNAS 95:10902–10907.

    https://doi.org/10.1073/pnas.95.18.10902

    • Google Scholar
50. Conference

    1. Minka T

    (2000)

    Automatic choice of dimensionality for pca

    Advances in Neural Information Processing Systems, 13.

    • Google Scholar
51. 1. Morgenroth E
    2. Vilaclara L
    3. Muszynski M
    4. Gaviria J
    5. Vuilleumier P
    6. Van De Ville D

    (2023) Probing neurodynamics of experienced emotions-a Hitchhiker’s guide to film fMRI

    Social Cognitive and Affective Neuroscience 18:nsad063.

    https://doi.org/10.1093/scan/nsad063

    • PubMed
    • Google Scholar
52. 1. Murphy FC
    2. Nimmo-smith I
    3. Lawrence AD

    (2003) Functional neuroanatomy of emotions: a meta-analysis

    Cognitive, Affective, & Behavioral Neuroscience 3:207–233.

    https://doi.org/10.3758/CABN.3.3.207

    • Google Scholar
53. 1. Ochsner KN
    2. Bunge SA
    3. Gross JJ
    4. Gabrieli JDE

    (2002) Rethinking feelings: an FMRI study of the cognitive regulation of emotion

    Journal of Cognitive Neuroscience 14:1215–1229.

    https://doi.org/10.1162/089892902760807212

    • PubMed
    • Google Scholar
54. 1. Ochsner KN
    2. Gross JJ

    (2008) Cognitive emotion regulation: insights from social cognitive and affective neuroscience

    Current Directions in Psychological Science 17:153–158.

    https://doi.org/10.1111/j.1467-8721.2008.00566.x

    • PubMed
    • Google Scholar
55. 1. Ogawa S
    2. Tank DW
    3. Menon R
    4. Ellermann JM
    5. Kim SG
    6. Merkle H
    7. Ugurbil K

    (1992) Intrinsic signal changes accompanying sensory stimulation: functional brain mapping with magnetic resonance imaging

    PNAS 89:5951–5955.

    https://doi.org/10.1073/pnas.89.13.5951

    • Google Scholar
56. 1. Olman CA
    2. Davachi L
    3. Inati S

    (2009) Distortion and signal loss in medial temporal lobe

    PLOS ONE 4:e8160.

    https://doi.org/10.1371/journal.pone.0008160

    • PubMed
    • Google Scholar
57. 1. Palomero-Gallagher N
    2. Amunts K

    (2022) A short review on emotion processing: a lateralized network of neuronal networks

    Brain Structure & Function 227:673–684.

    https://doi.org/10.1007/s00429-021-02331-7

    • PubMed
    • Google Scholar
58. 1. Parker D
    2. Liu X
    3. Razlighi QR

    (2017) Optimal slice timing correction and its interaction with fMRI parameters and artifacts

    Medical Image Analysis 35:434–445.

    https://doi.org/10.1016/j.media.2016.08.006

    • PubMed
    • Google Scholar
59. 1. Pessoa L

    (2017) A network model of the emotional brain

    Trends in Cognitive Sciences 21:357–371.

    https://doi.org/10.1016/j.tics.2017.03.002

    • Google Scholar
60. 1. Pessoa L
    2. McMenamin B

    (2017) Dynamic networks in the emotional brain

    The Neuroscientist 23:383–396.

    https://doi.org/10.1177/1073858416671936

    • PubMed
    • Google Scholar
61. 1. Phan KL
    2. Wager T
    3. Taylor SF
    4. Liberzon I

    (2002) Functional neuroanatomy of emotion: a meta-analysis of emotion activation studies in PET and fMRI

    NeuroImage 16:331–348.

    https://doi.org/10.1006/nimg.2002.1087

    • PubMed
    • Google Scholar
62. 1. Riedel MC
    2. Yanes JA
    3. Ray KL
    4. Eickhoff SB
    5. Fox PT
    6. Sutherland MT
    7. Laird AR

    (2018) Dissociable meta-analytic brain networks contribute to coordinated emotional processing

    Human Brain Mapping 39:2514–2531.

    https://doi.org/10.1002/hbm.24018

    • PubMed
    • Google Scholar
63. 1. Rosen AFG
    2. Roalf DR
    3. Ruparel K
    4. Blake J
    5. Seelaus K
    6. Villa LP
    7. Ciric R
    8. Cook PA
    9. Davatzikos C
    10. Elliott MA
    11. Garcia de La Garza A
    12. Gennatas ED
    13. Quarmley M
    14. Schmitt JE
    15. Shinohara RT
    16. Tisdall MD
    17. Craddock RC
    18. Gur RE
    19. Gur RC
    20. Satterthwaite TD

    (2018) Quantitative assessment of structural image quality

    NeuroImage 169:407–418.

    https://doi.org/10.1016/j.neuroimage.2017.12.059

    • PubMed
    • Google Scholar
64. 1. Saarimäki H

    (2021) Naturalistic stimuli in affective neuroimaging: a review

    Frontiers in Human Neuroscience 15:675068.

    https://doi.org/10.3389/fnhum.2021.675068

    • PubMed
    • Google Scholar
65. 1. Satpute AB
    2. Lindquist KA

    (2019) The default mode network’s role in discrete emotion

    Trends in Cognitive Sciences 23:851–864.

    https://doi.org/10.1016/j.tics.2019.07.003

    • Google Scholar
66. 1. Seeley WW

    (2019) The salience network: a neural system for perceiving and responding to homeostatic demands

    The Journal of Neuroscience 39:9878–9882.

    https://doi.org/10.1523/JNEUROSCI.1138-17.2019

    • PubMed
    • Google Scholar
67. 1. Sladky R
    2. Friston KJ
    3. Tröstl J
    4. Cunnington R
    5. Moser E
    6. Windischberger C

    (2011) Slice-timing effects and their correction in functional MRI

    NeuroImage 58:588–594.

    https://doi.org/10.1016/j.neuroimage.2011.06.078

    • PubMed
    • Google Scholar
68. 1. Smith SM
    2. Fox PT
    3. Miller KL
    4. Glahn DC
    5. Fox PM
    6. Mackay CE
    7. Filippini N
    8. Watkins KE
    9. Toro R
    10. Laird AR
    11. Beckmann CF

    (2009) Correspondence of the brain’s functional architecture during activation and rest

    PNAS 106:13040–13045.

    https://doi.org/10.1073/pnas.0905267106

    • Google Scholar
69. 1. Sonkusare S
    2. Breakspear M
    3. Guo C

    (2019) Naturalistic stimuli in neuroscience: critically acclaimed

    Trends in Cognitive Sciences 23:699–714.

    https://doi.org/10.1016/j.tics.2019.05.004

    • Google Scholar
70. 1. Vemuri K
    2. Surampudi BR

    (2015) Evidence of stimulus correlated empathy modes--Group ICA of fMRI data

    Brain and Cognition 94:32–43.

    https://doi.org/10.1016/j.bandc.2014.12.006

    • PubMed
    • Google Scholar
71. 1. Visser M
    2. Jefferies E
    3. Embleton KV
    4. Lambon Ralph MA

    (2012) Both the middle temporal gyrus and the ventral anterior temporal area are crucial for multimodal semantic processing: distortion-corrected fMRI evidence for a double gradient of information convergence in the temporal lobes

    Journal of Cognitive Neuroscience 24:1766–1778.

    https://doi.org/10.1162/jocn_a_00244

    • PubMed
    • Google Scholar
72. 1. Wallenwein LA
    2. Schmidt SNL
    3. Hass J
    4. Mier D

    (2024) Cross-modal decoding of emotional expressions in fMRI-Cross-session and cross-sample replication

    Imaging Neuroscience 2:1–15.

    https://doi.org/10.1162/imag_a_00289

    • PubMed
    • Google Scholar
73. 1. Waugh CE
    2. Shing EZ
    3. Avery BM

    (2015) Temporal dynamics of emotional processing in the brain

    Emotion Review 7:323–329.

    https://doi.org/10.1177/1754073915590615

    • Google Scholar
74. 1. Westermann R
    2. Spies K
    3. Stahl G
    4. Hesse FW

    (1996) Relative effectiveness and validity of mood induction procedures: a meta-analysis

    European Journal of Social Psychology 26:557–580.

    https://doi.org/10.1002/(SICI)1099-0992(199607)26:4<557::AID-EJSP769>3.0.CO;2-4

    • Google Scholar
75. 1. Windischberger C
    2. Cunnington R
    3. Lamm C
    4. Lanzenberger R
    5. Langenberger H
    6. Deecke L
    7. Bauer H
    8. Moser E

    (2008) Time-resolved analysis of fMRI signal changes using Brain Activation Movies

    Journal of Neuroscience Methods 169:222–230.

    https://doi.org/10.1016/j.jneumeth.2007.11.033

    • PubMed
    • Google Scholar
76. 1. Xu S
    2. Zhang Z
    3. Li L
    4. Zhou Y
    5. Lin D
    6. Zhang M
    7. Zhang L
    8. Huang G
    9. Liu X
    10. Becker B
    11. Liang Z

    (2023) Functional connectivity profiles of the default mode and visual networks reflect temporal accumulative effects of sustained naturalistic emotional experience

    NeuroImage 269:119941.

    https://doi.org/10.1016/j.neuroimage.2023.119941

    • PubMed
    • Google Scholar
77. 1. Yeo BTT
    2. Krienen FM
    3. Sepulcre J
    4. Sabuncu MR
    5. Lashkari D
    6. Hollinshead M
    7. Roffman JL
    8. Smoller JW
    9. Zöllei L
    10. Polimeni JR
    11. Fischl B
    12. Liu H
    13. Buckner RL

    (2011) The organization of the human cerebral cortex estimated by intrinsic functional connectivity

    Journal of Neurophysiology 106:1125–1165.

    https://doi.org/10.1152/jn.00338.2011

    • PubMed
    • Google Scholar
78. Book

    1. Zemeckis R
    2. Hanks T
    3. Wright R
    4. Sinise G
    5. Williamson M
    6. Field S
    7. Groom W

    (1994)

    Forrest Gump

    Paramount Pictures.

    • Google Scholar
79. 1. Zhou F
    2. Zhao W
    3. Qi Z
    4. Geng Y
    5. Yao S
    6. Kendrick KM
    7. Wager TD
    8. Becker B

    (2021) A distributed fMRI-based signature for the subjective experience of fear

    Nature Communications 12:6643.

    https://doi.org/10.1038/s41467-021-26977-3

    • PubMed
    • Google Scholar

Author details

1. Niels Janssen

   1. Department of Psychology, Universidad de la Laguna, La Laguna, Spain
   2. Institute of Biomedical Technologies, Universidad de La Laguna, La Laguna, Spain
   3. Institute of Neurosciences, Universidad de la Laguna, La Laguna, Spain

   Contribution

   Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation, Visualization, Methodology, Writing – original draft, Project administration, Writing – review and editing

   For correspondence

   njanssen@ull.es

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0001-9770-4458

2. Uriel KA Elvira

   Department of Psychology, Universidad de la Laguna, La Laguna, Spain

   Contribution

   Formal analysis, Methodology, Writing – review and editing

   Competing interests

   No competing interests declared

3. Joost Janssen

   1. Department of Child and Adolescent Psychiatry, Institute of Psychiatry and Mental Health, Hospital General Universitario Gregorio Marañón, Madrid, Spain
   2. Ciber del Área de Salud Mental, Instituto de Investigación Sanitaria Gregorio Marañón, Madrid, Spain
   3. Department of Psychiatry, UMCU Brain Center, University Medical Center Utrecht, Utrecht, Netherlands

   Contribution

   Methodology, Writing – review and editing

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0001-7613-2067

4. Theo GM van Erp

   1. Clinical Translational Neuroscience Laboratory, Department of Psychiatry and Human Behavior, University of California, Irvine, Irvine, United States
   2. Center for the Neurobiology of Learning and Memory, University of California, Irvine, Irvine, United States

   Contribution

   Methodology, Writing – review and editing

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0002-2465-2797

• 1,328

  views

• 88

  downloads

• 0

  citations

Views, downloads and citations are aggregated across all versions of this paper published by eLife.