Thanks to the generative nature of the scheme, the segmentation method is robust to changes in MRI contrast or acquisition hardware. Using cross validation, we show that the algorithm can segment the structures in previously unseen T1 and FLAIR scans with great accuracy (mean error under 1 mm) and robustness (no failures in 383 scans including 168 AD cases).
Voxel-based morphometry at ultra-high fields: A comparison of 7T and 3T MRI data ($)→
/My take: hard to know which is "better".
Miniaturized optical neuroimaging in unrestrained animals ($)→
/In this review, we highlight recent advances in the fabrication, characterization and application of miniaturized head-mounted optical neuroimaging systems pioneered by innovative investigators from a wide array of disciplines. We broadly classify these systems into those based on exogenous contrast agents, such as single- and two-photon microscopy systems; and those based on endogenous contrast mechanisms, such as multispectral or laser speckle contrast imaging systems. Finally, we conclude with a discussion of the strengths and weaknesses of these approaches along with a perspective on the future of this exciting new frontier in neuroimaging.
Dorothy Bishop on what a new face of scientific publishing might look like→
/No traditional journals; it's all "consensual communication". Open science built in (emphasis added):
- The study is then completed, written up in full, and reviewed by the editor. Provided the authors have followed the protocol, no further review is required. The final version is deposited with the original preprint, together with the data, materials and analysis scripts.
The Motherhood+PublicPower Index→
/I haven't looked at the details of the index, but I like the idea.
When approximately 40 percent of the US population are mothers, how can we be satisfied with just 14 percent representation in the halls of power? And with more than 3 out of every 4 of the most powerful positions held by fathers, clearly having children need not act as a barrier to public influence.
Measuring things can really help.
Online course for visualization in R from beginner to advanced→
/Looks nice!
You work through tutorials and pause to learn important concepts so that you can get into more advanced visualization. Exercises and suggested reading at the end of each week help you practice what you learn. By the end of the course, you will:
- Know how to work with real data
- Be able to explore data for analysis
- Create custom graphics fit for presentation
Scientific publishers' confidentiality clauses keep journal costs secret (!)→
/I did not know this. Crazy.
Controversially (and maybe illegally), when negotiating contracts with libraries, publishers often insist on confidentiality clauses — so that librarians are not allowed to disclose how much they are paying. The result is an opaque market with no downward pressure on prices, hence the current outrageously high prices, which are rising much more quickly than inflation even as publishers’ costs shrink due to the transition to electronic publishing.
Nikos Logothetis quits primate research→
/Huge news. Logothetis' lab has been at the center of numerous aspects of nonhuman primate research, including understanding the neurophysiological underpinnings of the BOLD fMRI signal.
In a letter last week to fellow primate researchers, Logothetis cites a lack of support from colleagues and the wider scientific community as key factors in his decision. In particular, he says the Max Planck Society—and other organizations—should pursue criminal charges against the activists who target researchers.
Logothetis did not seem to feel supported by his colleagues:
Logothetis’s letter also faults his scientific colleagues in Tübingen for distancing themselves from the controversy. The neighboring Max Planck Institute for Developmental Biology posted a disclaimer on its website emphasizing that there are no monkeys at the institute, he notes, and colleagues at the nearby Hertie Institute for Clinical Brain Research refused to issue a declaration of support.
Logothetis is moving to rodents.
Scientists share their stories of sexism in publishing→
/Following up on #AddMaleAuthorGate, some of the many tales of sexism still rampant in academia.
One of Sang’s bad experiences came from a paper in which she tracked patterns of co-authorship in a leading journal in her field over the course of 10 years. She found that white men frequently publish together, whereas female and minority scientists are more often at the periphery of these networks.
“One of the reviewers argued that the reason there are so few women and black academics in the social networks is because the research they produced just isn’t good enough to get into the top journals — and the editor agreed,” Sang said.
Comparing tract-based spatial statistics (TBSS) and manual labeling as diffusion analysis methods to detect white matter abnormalities→
/FA values from manual ROI and TBSS were strongly correlated (r = 0.94, P < 0.0001). Both methods found decreased FA in most ROIs for HIE infants. There was no significant interaction between method and group, indicating a similar ability to detect FA differences (F(1,19) = 0.599, P = 0.449). Sensitivity (manual: 0.709, TBSS: 0.694, 95% CI [–0.136, 0.163], P = 0.856), specificity (manual and TBSS: 0.716, 95% CI [–0.133, 0.133], P = 1), and standard error (manual: 0.009, TBSS: 0.007) were comparable.
Update on NIA paylines for 2015→
/"Not quite final" update from NIA. Not quite as bad as I thought but not great.
The sobering news is that for most areas of research within NIA’s mission the payline will be at the 8th percentile, with 5th percentile for applications requesting $500,000 and over, and 12th and 14th percentiles, respectively, for new and early stage investigator R01s.
Reproducibility of neuroimaging analyses across operating systems→
/With FSL, most Dice coefficients between subcortical classifications obtained on different operating systems remain above 0.9, but values as low as 0.59 are observed. Independent component analyses (ICA) of fMRI data differ between operating systems in one third of the tested subjects, due to differences in motion correction.
Ouch.
A first step to correct these reproducibility issues would be to use more precise representations of floating-point numbers in the critical sections of the pipelines. The numerical stability of pipelines should also be reviewed.
Orthogonalization of regressors in fMRI models→
/Very nice and important discussion of regressor orthogonalization.
The occurrence of collinearity in fMRI-based GLMs (general linear models) may reduce power or produce unreliable parameter estimates. It is commonly believed that orthogonalizing collinear regressors in the model will solve this problem, and some software packages apply automatic orthogonalization. However, the effects of orthogonalization on the interpretation of the resulting parameter estimates is widely unappreciated or misunderstood. Here we discuss the nature and causes of collinearity in fMRI models, with a focus on the appropriate uses of orthogonalization. Special attention is given to how the two popular fMRI data analysis software packages, SPM and FSL, handle orthogonalization, and pitfalls that may be encountered in their usage. Strategies are discussed for reducing collinearity in fMRI designs and addressing their effects when they occur.
And as a bonus, examples in an IPython notebook. Very cool.
Reminder: There's more to improving science than p values→
/Good reminder that there is a lot more to improving the quality of science than p values.
P values are an easy target: being widely used, they are widely abused. But, in practice, deregulating statistical significance opens the door to even more ways to game statistics — intentionally or unintentionally — to get a result. Replacing P values with Bayes factors or another statistic is ultimately about choosing a different trade-off of true positives and false positives. Arguing about the P value is like focusing on a single misspelling, rather than on the faulty logic of a sentence.
How do you memorize an entire symphony?→
/Neuroscientists have found that musical memories can be preserved in the brain even when most other memories are lost. Memory loss, such as that experienced by amnesiacs, provides a window for neuroscientists to study how memory works.
One amnesiac patient, a German professional cellist, had such profound memory loss that he could not remember well-known facts about Germany, nor important details from his youth or adulthood. He had no memory of relatives and friends, except for his brother and his full-time care-giver.
However, he could still play the cello and sight-read music. Moreover, his memory for music from his past was just as good as that of his non-amnesiac colleagues. He could even learn to recognise new music (but not new faces or objects).
What should science look like?→
/Reimagining of what science should be like from Björn Brembs. Lots of interesting ideas here. I love the idea of being able to explore published data:
I would also be able to double-click on any figure to have a look at other aspects of the data, e.g. different intervals, different intersections, different sub-plots. I’d be able to use the raw data associated with the publication to plot virtually any graph from the data, not just those the others offer me as a static image, as today.
...which requires authors to share their data and link it to publication:
As an author, I want my data to be taken care of by by institution: I want to install their client to make sure every piece of data I put on my ‘data’ drive will automatically be placed in a data repository with unique identifiers.
Bonus points for being posted/published on The Winnower.
Beyond bar and line graphs: Time for a new data presentation paradigm→
/Different data can lead to the same summary statistics - hence, authors should try to show as much data as possible (e.g., scatter plots rather than bar graphs of means). Good advice.
R01 teams and grantee age trends at NIA→
/Interesting (and possibly depressing) post from Robin Barr (NIA) on funding trends.
The average age of first-time R01 funded investigators who have PhDs remains 42 even after seven years of policies at NIH to increase the numbers of new and early-stage investigators. And, over the same interval, age has continued to increase for first-time R01-funded MDs and MD-PhDs, despite the policies we have in place.
This comment on the number of investigators on an R01 caught my eye:
How many investigators does it take to write an R01? I looked at the 100 top-scoring R01 applications across NIH in January 2015 and compared them to a similar set from January 2005. R01 applications have been bulking up! In 2005, more of the top scoring applications had a single principal investigator listed as the faculty on that application—just Professor X and the students and postdocs—than had two faculty, or three faculty or any other number. By 2015, Professor X needed more help. Now, three faculty is the most common number of faculty members on an application. By 2015, the “average” top-scoring R01 at NIH had more than four faculty listed as participating on it.
I don't think it's that people need more help writing R01s; it's that funding is tight and people are scrambling to try to cover salary support.
New paper describing @neurovault→
/Frontiers paper describing neurovault, the very cool data sharing service and viewer for MRI (etc.) results.
Increasing diversity at your conference→
/Great compilation of advice from Ashe Dreyden. Tech-centered but a lot of it is very applicable to academice conferences, too.
The easiest way to get feedback on your efforts is to publically state what you've tried and ask for contructive criticism. Be transparent and truthful. I've seen many conferences write blog posts about what they've done to address the issue of the lack of diversity and the positive or negative results that they ended up with. This is important for a few reasons: it signals that this is important to you and that you are open to more ideas as well as letting people within marginalized groups know that you are considering their needs and the reality of their situations.
