On this page you can find the answers from my (unsuccessful) application to become a Mozilla Fellow for Science in 2015.
The application was submitted on August 14th at 7:01pm while on holiday in Edinburgh! I had some delicious cocktails that evening .
This work is licensed under a Creative Commons Attribution 4.0 International License which means you can distribute, remix, tweak, and build upon these answers as long as you credit me for the original creation.
What research fields are you in?
Longitudinal neuroimaging, brain development, adolescent psychopathology, MRI physics, network analysis
What is your research focus?
I aim to integrate multiple neuroimaging techniques of brain development during childhood and adolescence to predict individual differences in susceptibility to mental health problems and response to interventions. I use graph theory to investigate changes in MRI measures of brain structure and function, and the biological mechanisms underlying brain networks.
Describe to us your current research team
The Neuroscience in Psychiatry Network is a collaboration between the University of Cambridge and University College London, funded by a £5 million Wellcome Trust Strategy Award. The team comprises 5 principal investigators, 2 project managers, 10 postdoctoral fellows, 5 graduate students, and 12 administrative staff situated across 7 geographic locations.
During my PhD I was mentored and inspired by exceptional scientists and computer programmers. In my 3 years at the University of Cambridge, I have picked up their baton and helped collaborators, researchers in my department, and over 50 members of the Cambridge Women in Technology meetup group learn to code. I am incensed when I witness (particularly) women being instructed to rely on others for their mathematical, computation or statistical analyses and am dedicated to empowering my peers to produce clinically relevant, reproducible and rigorous results through continued education and interdisciplinary collaboration.
As a physicist I was taught, and have taught others, to always provide proof of my logical inductions and to show my working. I follow these principles today by ensuring my neuroimaging analyses can be independently reproduced: pre-processing is scripted, statistical calculations are performed in R or Python and figures are created from the original data using the open source packages in Python. My code is commented and publicly available at www.GitHub.com/KirstieJane.
The field of neuroimaging has experienced a crisis of false positives and misleading statistical thresholding, but there is light at the end of the tunnel. The Mozilla Science Lab, and others, have inspired a critical mass of researchers who understand the benefits of meta-analyses, publishing null-results, pre-registration, sharing data and collaborative working. I relish this opportunity to extend and expand all of our efforts as a champion for open science and ambassador for the open web. Thank you for your consideration.
Yours sincerely, Kirstie Whitaker
Describe to us how open science advances your research
Psychiatric disorders are responsible for 184 million disability-adjusted life years worldwide (Whiteford et al, 2013) and there is strong evidence that environmental and genetic effects affect the structure and function of the developing brain, which in turn leads to individual differences in each person’s susceptibility to mental health disorders. However, neuroimaging studies are often underpowered and are always expensive and statistically noisy. When these data sets are published alongside their findings everyone benefits: researchers investigating similar questions may validate their findings in an independent cohort, computational modellers may test new methods, and analytic errors can be caught and corrected.
What work are you currently involved in that’s relevant to being an open science leader? Are you leading any projects related to open science?
NSPN is made up of talented statisticians, clinicians, physicists, psychologists and epidemiologists, and every member is passionate about producing ground-breaking interdisciplinary research. I have taken responsibility for designing and implementing MRI processing and quality assurance pipelines, and for providing documentation and processed values to other members of the team. I am passionate about empowering my colleagues and explaining code in my GitHub repositories rather than encouraging a “black box” analysis strategy. I coordinate and lead regular meetings for all early career researchers within the network in order to facilitate collaboration and the exchange of new analysis hypotheses, techniques and interpretations.
How would this fellowship accelerate your work?
This fellowship would provide me with the networking and technical support required to integrate the different data types collected within NSPN and to share them online while maintaining strict levels of participant anonymity. I will be faster and more efficient in my analyses, as will our internal and external collaborators.
What do you see as the opportunities for impact around open research at your university? Could you leverage this opportunity in a potential project?
The University of Cambridge has been a leader in the natural sciences since its foundation in 1209 and remains committed to its position as an international role model for academic research. Financial and educational resources to facilitate data sharing are available (cf. www.data.cam.ac.uk) but I will develop personalised curricula to ensure their visibility, uptake and understanding. In sharing data from the multi-disciplinary Neuroscience in Psychiatry Network I will bring external collaborations to the university, increase the impact of our work on the development of mental health disorders, and serve as a flagship example of best practise in open science.
What do you think needs to change most immediately in the system?
Raw data, all analysis code and the instructions to replicate all figures need to be made available to reviewers with all submitted journal articles. It is simply not enough to say what one did or found, the data and code must be available for independent verification.
What project in the field do you find most inspiring to further science and the web?
I am most inspired by the Neurosynth project (www.neurosynth.org): an online platform for large-scale meta-analyses of functional MRI data through text mining published articles. You can explore associations between regional brain activations and each of over 3000 terms, or decode your own results in real time by reversing the process.
Why is the open web important to you?
The open web is the great equaliser. It doesn’t matter where you come from, your gender, sexual orientation or race: the open web allows you to meet peers, mentors, role models and friends. Academia is competitive and, sometimes, dangerously isolating. The mental health challenges for highly educated “trainees” is often not appreciated by those in power and the “crushing self-doubt” of the imposter complex pervades our halls. The open web allows science to advance by sharing our deeper understandings (rather than only the novel or unexpected results) and by supporting the real human beings who undertake those analyses every day.