Pre-Registration: Response Priming by Colour in FMRI, EEG, and Behavior: A Multimethod Study. Part 2: FMRI

RPTU University of Kaiserslautern-Landau Preregistration “Response Priming by Color in FMRI, EEG, and Behavior: A Multimethod Study. Part 2: FMRI” Time Point of Pre-Registration This analysis plan is pre-registered prior to any analysis of data designed to answer the research questions. So far, the data have been accessed for purposes of data management and matching of MRI and behavioral data. Theoretical Background In this experiment, we investigated neural correlates of feed-forward response priming. In particular, we were interested in the neural evolution of priming effects across different stimulus-onset-asynchronies (SOAs) as predicted by the Rapid Chase Theory of Response Priming (Schmidt, Niehaus, & Nagel, 2006) and accumulator models of response priming (e.g., Schmidt & Schmidt, 2018). These theories assume that, initially, solely the prime activates the response associated with it, until the target stimulus is presented and drives the response henceforth. One corollary and explicit prediction is an increase in priming effects with increasing SOAs until about 100 ms, since the activation of an erroneous response by an incongruent prime will be larger with larger SOAs and requires more time to be compensated for by correct response activation by the target as soon as it has been presented. The resulting increase in response times for incongruent trials, thus, increases the priming effect (defined as RTs for incongruent – RTs for congruent trials). Considering the location of effects in related previous fMRI research (Krüger et al., 2013; Lee et al., 2020; Zhang et al., 2017), we were particularly interested in effects in the motor cortex, inferior frontal lobe, insula, dorsal anterior cingulate cortex (dACC), and supplementary motor area (SMA). In these areas, we aimed at identifying (A) brain regions showing a general response priming effect (incongruent > congruent; irrespective of SOA or response laterality), (B) brain regions showing lateralised response priming effects (e.g., in motor cortex), requiring separate analyses for “left” and “right” responses, (C) brain regions in which the priming effect (incongruent > congruent) increased with increasing SOA, again separately for “left” and “right” responses in order detect potential lateralised effects (e.g., motor cortex). In addition, we were interested in (D) whether longer response times in incongruent trials are also associated with enhanced activation in the motor area opposite to the final, correct response (i.e., increased left hemispheric motor activation for slower, but eventually correct “left” responses). Methods Task description The task description can be found in detail in a separate pre-registration focusing on behavioral data analysis (Schmidt, 2025). In short, we employed a responses priming paradigm (Vorberg et al., 2003). Each trial consisted of a fixation period (duration: 767 ms), a flanker prime (two same-colored circles located left and right from the center, either both red or green; duration 33.4 ms), followed by a variable SOA (33, 83, 117, or 150 ms) during which a blank screen was presented, and, finally, the target stimulus (a single centrally located circle, either red or green; duration: 1017 ms), whose presentation duration also served as the response window. Participants were instructed to indicate the color or the target stimulus. A variable intertrial interval separated trials (duration jittered using optseq (https://surfer.nmr.mgh.harvard.edu/optseq/; Dale, 1999); M = 5,025 ms, SD = 1,567 ms, range: 2,000 to 8,000 ms). Participants completed 400 trials (50 trials per condition congruency [congruent, incongruent] x SOA [33, 83, 117, 150 ms]), split into 4 blocks à 100 trials. Assignment of keys to response options (“left” = “red” and “right” = “green” or vice versa) was counterbalanced across participants. Data exclusions On participant level, individuals will be excluded if they have incomplete data sets (e.g., due to technical issues or discontinuation of the study), or show low task performance (mean error rate or reaction time two standard deviations below the sample mean in either the congruent or incongruent condition). Individuals will also be excluded if they moved more than 4 mm in any direction along any of the three axes (translation or rotation). Individuals excluded from fMRI analysis due to movement during the task will still be included in the behavioral data analysis. On trial level, trials with very short (< 100 ms) or very long response times (exceeding target duration of 1,017 ms) or with a prime presentation duration of 50 ms instead of 33.4 ms (due to technical error) will be excluded from behavioral data analysis and modelled as part of the error regressor in the fMRI analysis. Participants 31 participants took part in this fMRI study. Four of these participants did not complete data collection, six participants have incomplete data sets due to technical issues during scanning, leaving 21 participants before exclusion due to performance or motion. MRI image acquisition All images were acquired with a 3-Tesla MR scanner (Siemens Magnetom Vida, Erlangen, Germany) located at the Westpfalz-Klinikum, Kaiserslautern, using a 20channel head coil. Anatomical data were acquired with an MP-RAGE sequence (320 x 320 matrix, 1 mm isometric voxel size, 224 sagittal slices, FoV 305 mm, TR 2250 ms, TE 2 ms). Functional MRI data were acquired using a gradient-echo echo-planar imaging sequence (TR 2000 ms, TE 30 ms, FoV 192 mm, 70° flip angle, 64 x 64 matrix). Each image consisted of 33 slices (3.6 mm thickness, 20 % interslice gap), which were acquired in interleaved ascending order, an anterior-posterior phase-encoding direction and a parallel imaging factor of 2. Analysis of behavioral data from MRI experiment Behavioral data analyses have been pre-registered separately (Schmidt, 2025). MRI analysis All MRI data will be processed and analysed using SPM12 (Statistical Parametric Mapping, Wellcome Department of Cognitive Neurology, London, UK). Pre-processing Since the MRI sequence was longer than the task and had to be manually stopped once the participants had finished the task, the final image of each run will be deleted if it is incomplete (i.e., contains less slices than 33 slices). The functional images will be first slice-time corrected to the 17th slice, and spatially realigned to the mean image of the series (estimation interpolation: 2nd degree B-spline; writing interpolation: 4th degree B-Spline). Movement estimation plots will be visually screened for participants who moved more than 4 mm along any of the three axes (either in translation or rotation). The functional images will then be co-registered to the individual anatomical image using the normalized mutual information criterion. Anatomical images will be nonlinearly transformed to the MNI 152 space (Montreal Neurological Institute) and resliced with a spatial resolution of 1 mm isotropic voxel size (interpolation: 5th degree B-Spline). Subsequently, the transformation will be applied to the functional volumes, which will be resliced with a spatial resolution of 2 mm isotropic voxel size (interpolation: 5th degree B-Spline). Finally, images will be smoothed with a Gaussian kernel of 8 mm full width half maximum. First-level GLM First-level analysis of individual data will use a general linear model (GLM) based on an informed basis set consisting of the canonical hemodynamic response function. Stimulus onsets that will be entered as event-related regressors into the GLM include invalid trials (containing trials with incorrect or missing responses, trials with very fast responses < 100 ms, and trials with timing issue in prime timing, see above) as well as task regressors for each combination of the conditions ‘congruency’ (congruent, incongruent), ‘response laterality (left, right), and ‘SOA’ (33, 83, 117, 150 ms), all entered with a duration of 0 seconds. Regressors will be concatenated across runs. Six motion regressors (rotation and translation in x, y, and z direction) and an intercept for each run will be added as nuisance regressors. Low frequency drifts in the signal will be accounted for by a high-pass filter (128 Hz). A global approximate AR(1) autocorrelation model will be used. Individual contrasts will be set up. These contrasts will be stated in the following, given the following order of individual betas: incongruent_33_left, incongruent_83_left, incongruent_117_left, incongruent_150_left, incongruent_33_right, incongruent_83_right, incongruent_117_right, incongruent_150_right, congruent_33_left, congruent_83_left, congruent_117_left, congruent_150_left, congruent_33_right, congruent_83_right, congruent_117_right, congruent_150_right Research Question (A) – Contrast 1: One t-contrast will be coding for a general congruency/priming effect (incongruent vs. congruent) across all correct responses, irrespective of laterality (left or right): [1 1 1 1 1 1 1 1 -1 -1 -1 -1 -1 -1 -1 -1] Research Question (B) – Contrast 2 & Contrast 3: Two t-contrasts will code for the congruency/priming effects separately for left vs. right responses: Left: [1 1 1 1 0 0 0 0 -1 -1 -1 -1 0 0 0 0] Right: [0 0 0 0 1 1 1 1 0 0 0 0 -1 -1 -1 -1] Research Question (C) - Contrasts 4 to 6 & 7 to 9: Three t-contrasts will, finally, code for an interaction of the congruency/priming effect and SOA. Each contrast will test for a difference in the priming effect between two adjacent SOAs (33 ms vs. 83 ms, 83 ms vs. 117 ms, 117 ms vs. 150 ms). Again, contrasts will be specified separately for left and right responses: Left: Contrast 4: [-1 1 0 0 0 0 0 0 1 -1 0 0 0 0 0 0] Contrast 5: [0 -1 1 0 0 0 0 0 0 1 -1 0 0 0 0 0] Contrast 6: [0 0 -1 1 0 0 0 0 0 0 1 -1 0 0 0 0] Right: Contrast 7: [0 0 0 0 -1 1 0 0 0 0 0 0 1 -1 0 0] Contrast 8: [0 0 0 0 0 -1 1 0 0 0 0 0 0 1 -1 0] Contrast 9: [0 0 0 0 0 0 -1 1 0 0 0 0 0 0 1 -1] In order to answer the question of whether longer reaction times in the incongruent condition are associated with enhanced activation in the motor area opposite to the final, correct response (i.e., increased left hemispheric motor activation for slower, but eventually correct “left” responses), (Research Question D) we will set up an additional first-level model for each participant. Each model will include the same regressors as described above, however, a parametric modulator containing the reaction time of the respective trial will be added to the incongruent regressors of the one SOA, for which the largest behavioral priming effect in RTs (across participants) was observed. The parametric modulator will be applied to both incongruent regressors for “left” as well as “right” responses. Individual t-contrasts will be created, coding 1 for the respectively parametric modulator, and 0 for every other regressor. Second-level GLM Second-level mass univariate analysis will use one-sample t-tests with the individual contrast maps (see above) as criterion, for testing contrasts 1 to 3 (Research Questions A and B), separately. The intercept of these second-level models will then be tested for positive and negative effects. For contrasts 4 to 9 (Research Question C), two secondlevel repeated-measures one-way ANOVAs will be set up (one for “left” responses [including contrasts 4 to 6], and one for “right” responses [including contrasts 7 to 9]), containing one factor with the three SOA comparisons as levels (i.e., 33 ms vs. 83 ms, 83 ms vs. 117 ms, 117 ms vs. 150 ms). After estimation, an F-contrast will test for any effects across the three SOA comparisons ([1 0 0; 0 1 0; 0 0 1]). If the F-test yields significant results, each of the three t-contrasts/SOA comparisons will be entered into a separate second-level one-sample t-test and tested for both positive and negative effects. Finally, to identify brain regions in which the congruency-effect monotonically increases across all SOA levels (i.e., to test whether the three SOA differences 33 vs. 83, 83 vs. 117, 117 vs. 150 ms, are all significant), a conjunction analysis including all three second-level t-contrasts will be conducted. In order to answer the question of whether longer reaction times in the incongruent condition are associated with enhanced activation in the motor area opposite to the final, correct response (Research Question D), the t-contrasts for the parametric modulators will be entered as criterion into two separate one-sample t-tests, one for left and one for right responses. The intercept will then be tested for both positive and negative effects for all t-tests. In all second-level analyses, the color-key-assignment (left = red, right = green, or vice versa) for each participant and participant sex (coded -0.5 and 0.5) will be entered as covariates, and all analyses will be thresholded at pvoxel < .001 (uncorrected), and a cluster size threshold of pcluster < .05 (FWE-corrected). Since we expect findings to be located in a specific set of brain regions (see above), we will first apply a combined explicit mask of these brain regions. These anatomical masks will be taken from the AAL3 (Rolls et al., 2020), and will include the following AAL3-ROIs, all bilaterally: Precentral, Frontal_Inf_Oper, Frontal_Inf_Tri, Frontal_Inf_Orb, Insula, Frontal_Sup_Med, and Supp_Motor_Area. Subsequently, we will also conduct whole-brain analyses. Explorative Analyses In addition, we will explore effects in the parietal lobe. References Dale, A. M. (1999). Optimal experimental design for event-related fMRI. Human Brain Mapping, 8(2–3), 109–114. https://doi.org/10.1002/(SICI)1097-0193(1999)8:2/3<109::AIDHBM7>3.0.CO;2-W Krüger, D., Klapötke, S., Bode, S., & Mattler, U. (2013). Neural correlates of control operations in inverse priming with relevant and irrelevant masks. NeuroImage, 64, 197–208. https://doi.org/10.1016/j.neuroimage.2012.09.018 Lee, S.-M., Henson, R. N., & Lin, C.-Y. (2020). Neural Correlates of Repetition Priming: A Coordinate-Based Meta-Analysis of fMRI Studies. Frontiers in Human Neuroscience, 14. https://doi.org/10.3389/fnhum.2020.565114 Rolls, E. T., Huang, C.-C., Lin, C.-P., Feng, J., & Joliot, M. (2020). Automated anatomical labelling atlas 3. NeuroImage, 206, 116189. https://doi.org/10.1016/j.neuroimage.2019.116189 Schmidt, T. (2025). Preregistration: „Response priming by color in fMRI, EEG, and behavior: a multimethod study. Part 1: Behavior“. https://zenodo.org/records/15481264 Vorberg, D., Mattler, U., Heinecke, A., Schmidt, T., & Schwarzbach, J. (2003). Different time courses for visual perception and action priming. Proceedings of the National Academy of Sciences, 100(10), 6275–6280. https://doi.org/10.1073/pnas.0931489100 Zhang, R., Geng, X., & Lee, T. M. C. (2017). Large-scale functional neural network correlates of response inhibition: An fMRI meta-analysis. Brain Structure and Function, 222(9), 3973– 3990. https://doi.org/10.1007/s00429-017-1443-x