Main content



Loading wiki pages...

Wiki Version:
# Enhancing effect of caffeine on mnemonic discrimination Frederik Aust & Christoph Stahl University of Cologne --- Mnemonic discrimination is the ability to discriminate among similar memories, which requires separable representations of similar information. The neurocomputational process that assumedly decorrelates representations during encoding and consolidation is referred to as pattern separation. Deficits in pattern separation contribute to age-related declines in mnemonic functioning, which has motivated the development of targeted interventions. We followed-up a recent report that one 200 mg-dose of caffeine administered post-study enhances mnemonic discrimination (Borota et al., 2014). To test whether the reported enhancements are an artifact of performance-impairing withdrawal symptoms in the control group, we did not restrict preexperimental caffeine consumption and statistically adjusted treatment effects for habitual caffeine consumption. We detected no effects of caffeine and nonsuperiority testing ruled out medium and large enhancements in both average (1200 mg per week) and low-consumers (50 mg per week). Our results raise doubts about a caffeine-mediated enhancement of mnemonic discrimination on two counts: If the effect exists, it (1) is substantially smaller than originally reported and (2) may reflect an offset of performance-impairing withdrawal symptoms rather than genuinely enhanced consolidation. We recommend that future studies employ an alternating exposure-abstinence protocol, use an active control group, and verify posttreatment caffeine abstinence via saliva or blood samples. --- This repository research products associated with the publication. We provide the experimental software, stimulus material that we are permitted to share in the `caps1/material` directory. The R Markdown files in the `paper` directory contain details about the analyses reported in the paper, as well as instructions on how to rerun the analysis to reproduce the results. With the help of the R package `rmarkdown` the file can be rendered into the manuscript in `PDF`-format. To do so render the file `paper/replication_borota.Rmd`. Note that rendering the manuscript as is requires unpublished demographic data (see Dataset description). The `caps1/results/data_raw/` directory contains all the collected raw data; merged and processed data files are provided in `caps1/results/data_processed/`. ## Publication (recommended citation) ... ## Dataset description Data were collected at the University of Cologne. | Study | Data collection period | | ------------ | ----------------------- | | Study 1 | 2014-05-13 - 2014-05-23 | Descriptions of the data collection methods are provided in the research report. Details about the data processing steps are available in the analysis file `R/aggregate_data.R` and the R Markdown file `paper/replication_borota.Rmd` used to generate the manuscript in `PDF`-format. To protect participant privacy we cannot share participants' demographic information publicly. Please contact the author to request a copy of these data. ## Software requirements ### Experimental software The experiment was programmed using [OpenSesame]( (Mathôt, Schreij, & Theeuwes, 2012). To run the experimental software copy the `experimental_software` folder, first run `study_phase.opensesame.tar.gz` and then `recognition_phase.opensesame.tar.gz`. ### Analyses Analyses were originally run on Ubuntu 16.04. We used R (Version 3.6.3; R Core Team, 2019) and the R-packages afex (Version 0.27.2; Singmann, Bolker, Westfall, & Aust, 2019), emmeans (Version 1.4.5; Lenth, 2019), mgcv (Version 1.8.31; Wood, 2011, 2003, 2004; Wood et al., 2016), papaja (Version; Aust & Barth, 2018), pwr (Version 1.3.0; Champely, 2018), shadowtext (Version 0.0.7; Yu, 2019), and TOSTER (Version 0.3.4; Lakens, 2017b) for all our analyses. To install the R-package `papaja`, which is required to reproduce the research report, please review the [installation instructions]( To rerun the analyses, render the file `paper/replication_borota.Rmd`. Note that rendering the manuscript as is requires unpublished demographic data (see Dataset description). ## Acknowledgements We thank Anika Preik, Marie Vogt, Christina Kessel, Ines Schwerdtfeger, Nele Wolf, and Sophie Maxim for help with the data collection. ## Licensing information | Material | License | | --------------------- | ------------------------------------------------------------ | | Experimental software | [MIT]( 2019 Frederik Aust, please cite the published article (if available) or prerint | | Data | [CC-BY-4.0](, please cite the published article (if available) or prerint | | Analysis code | [MIT]( 2019 Frederik Aust, please cite the published article (if available) or prerint | | Manuscript | [CC-BY-4.0](, please cite the published article (if available) or prerint | ## Contact Frederik Aust Herbert-Lewin-Str. 2 50931 Cologne Germany E-mail: ## References Mathôt, S., Schreij, D., & Theeuwes, J. (2012). OpenSesame: An open-source, graphical experiment builder for the social sciences. Behavior Research Methods, 44(2), 314–324.