As scientists, we are constantly striving for ways to optimize the design, implementation, and dissemination of our work. One of the most promising avenues towards reaching these goals is to embrace Open Science.
What is Open Science?
Open Science is a research approach that involves producing more transparent and reproducible research. Increasing transparency includes providing study information to fellow researchers as well as to the general public. This can include sharing data points and what kind of stimuli, methods, and analyses were used, as well as programming scripts that were used during any of these steps. Reproducibility refers to the reliability and consistency of findings. For example, examining how many studies that compared neural activation in response to images versus tones found significant activation differences in certain regions of the brain. You can increase the reproducibility of your study by organizing and preprocessing your data in a consistent manner and sharing these methods with others to ensure they could be able to reproduce similar results using your methods.
Why should we work towards increasing transparency and reproducibility?
The brain is a complex organ. There can be no one singular study that explains any one mechanism of the brain. Only by gathering information from numerous studies can we truly begin to understand some of the variances and complexities of human cognition. Consider the previous example about neural activation for images and tones. Of course, we would not expect all results to be identical since each study would likely use their own sets of images and tones, in their own set of participants, unless it was a replication experiment. For this reason, it can be incredibly useful for all the researchers to share their work to see which differences in stimuli or people are associated with differences in the brain. This helps us to more comprehensively understand the relationship between the brain and processing of these specific types of information.
Science is a complex and expensive process. As technology advances, we are constantly finding new ways to examine data. Open Science allows researchers to share new approaches. Towards examining data, Open Science enables fellow scientists to move forward without having to constantly re-invent the wheel. Furthermore, many of the imaging techniques that we use provide us with a multitude of data, to the extent where there can be data that is never actually published. These data often cost roughly $500-$2000 an hour to collect. By sharing our de-identified data (meaning we do not share any confidential or identifiable information), we can allow fellow researchers who might be interested in a subset of our data that may go unused to analyze these data, thereby maximizing the impact of our study while also potentially garnishing additional shared publications or citations of our data.
Science communication is essential. While there are many advantages to our ever increasingly connected world via the internet, there can also be widespread misinformation or misunderstandings. This can sometimes lead to mistrust of the scientific community as a whole. By sharing our data and methods with the general public, we can take a step forward towards building a bridge for people to understand our research. This especially makes sense given the fact that it is the public’s taxes that are paying for a large amount of the studies we conduct. This provides a mechanism for them to see what their tax dollars are going towards, and how these advances are changing the world. Of course, we also need more science communicators to get some of this information across as well.
In light of all these advantages, an increasing amount of publishers are now providing incentives or requirements for scientists to share their data. Furthermore, there has been a recent push towards pre-registering your study with publishers, wherein researchers describe their methods and hypotheses before conducting the actual research. This process guarantees the authors will have their research published regardless of their results. Given the traditional academic climate of “publish or perish”, which can lead to selective publishing of only positive results, pre-registering studies can enable scientists to publish null results, which can help us get a more extensive understanding of the data and allows for more reproducible research. Certainly, as with any large scale change, there are several caveats to this open science process. For example, scientists should never share identifiable or confidential information, nor should they share information that the participants have not already consented to sharing. Also, there is always the possibility that sharing data before you have actually analyzed the data could lead other researchers publishing articles on your data before you. So long as we are cognizant and plan ahead for these obstacles though, this movement has the potential to significantly advance the process of scientific research.
How do we implement an Open Science approach?
There are many ways to implement open science procedures. In our lab, we enact several procedures to increase transparency and reproducibility. For example, to enhance reproducibility, we use BIDS format to organize all our data and fmri-prep to pre-process all of our data, which ensures that we are organizing and preprocessing our data using consistent methods across labs. To increase transparency, our completed methodology and scripts that we use in the lab are openly available on our GitHub page. Furthermore, our de-identified data that have been consented to be shared will be uploaded to OpenNeuro, a data sharing platform for neuroimaging data.
In my opinion, the most effective way to increase the impact of our life’s work is to share it with others so we can all build on each other’s research. So what are you waiting for? Join the Open Science movement today!
— Jennifer Legault, Ph.D
Post-Doctoral Research Fellow in the QLAB