Tales of Neuro-Terror

By Neuroskeptic | November 3, 2013 9:21 am

At this time of year, people are fond of telling scary tales – generally involving ghosts, ghouls, and other frightening creatures. Neuroscientists have their own horror stories, however – more niche, perhaps, but no less terrifying.

Picture the scene: a group of PhD students are gathered around a flickering MRI console. The elder of the group (a fourth year) holds a torch pointing up at her face as she begins…


Many years ago, there was a student in the lab. No one even remembers his name – except the Professor, and he’s sworn never to speak of him.

Because you see, this student spent three years scanning volunteers with DTI for a study of white matter integrity . He found some good results, and he’d almost finished writing his thesis. He even got a paper accepted by Nature Neuroscience.

Until one cold autumn night, when the student was working late, running an analysis. He got an email from the FSL mailing list. A physicist reported a software bug that had been discovered on certain MRI scanners. The bug introduced artefacts into some kinds of data.

The student wasn’t too worried… that is, until he scrolled down the page, and saw an example of a corrupted scan.

He looked back up at the data he was analyzing… and he saw the exact same thing!

No one knows what happened to him after that. Some say he went crazy, and now lives in the air ducts at this very MRI center. Late at night, you might even hear his anguished cries… “I thought… the scans… were meant to look like that…”



This tale is loosely based on the true story of the ‘DTI vibration artefact‘. An artefact (pictured) was discovered to be affecting certain implementations of the Diffusion Tensor Imaging sequence. Lots of researchers had been using it, unwittingly collecting data, before someone revealed the problem.

That story ended relatively happily, because someone worked out how to clean up the data post hoc, but this still left the awkward question of how to interpret results that had already been published based on the affected data.

I know of a number of other neurohorror stories, along the same lines – stories of time, money, and effort that turn out to have been wasted due to some technical problem. These things can happen, I believe, because of the culture of treating raw data as secret and only sharing final results.

Most data quality issues would be spotted very quickly if anyone familiar with ‘good’ data were to inspect the affected scans, or vice versa; but all too often, neuroscientists never get the chance to see what other peoples’ raw data look like.

Please feel free to share any neurohorror stories of your own, in the comments.

ResearchBlogging.orgGallichan D, Scholz J, Bartsch A, Behrens TE, Robson MD, & Miller KL (2010). Addressing a systematic vibration artifact in diffusion-weighted MRI. Human Brain Mapping, 31 (2), 193-202 PMID: 19603408

  • Russ Poldrack

    As a postdoc I once played an April Fools joke, wherein I emailed the lab list to tell them that we had discovered that all of our data for the last year had an incorrect L-R orientation. That was a joke, but there are persistent rumors of papers that have been published where this kind of L/R confusion has been discovered after publication. Would not be surprising at all, given the persistent difficulty in establishing L/R orientation with certainty.

    • http://blogs.discovermagazine.com/neuroskeptic/ Neuroskeptic

      *makes a note of that prank for next April Fools*

  • Greig de Zubicaray

    The real horror lies in a failure to publish a corrigendum when an error in a published paper is discovered by its authors or when other researchers have brought the error to the authors’ attention. Without published corrigenda, there is no way of knowing which results/papers are erroneous (‘zombie papers’ anyone?), so I don’t see the happy ending to this story unless I assume all those prominent neuro labs that were first to publish with DTI had already corrected their data prior to publication or published corrigenda afterward…

  • Thomas Rhys Marshall

    I once had a diffusion protocol that was haunted by n/2 ghosts. Luckily they were discovered during piloting and exorcised.

    • http://blogs.discovermagazine.com/neuroskeptic/ Neuroskeptic




No brain. No gain.

About Neuroskeptic

Neuroskeptic is a British neuroscientist who takes a skeptical look at his own field, and beyond. His blog offers a look at the latest developments in neuroscience, psychiatry and psychology through a critical lens.


See More

@Neuro_Skeptic on Twitter


Discover's Newsletter

Sign up to get the latest science news delivered weekly right to your inbox!

Collapse bottom bar