FND and the replication crisis
Christoff Trüter
Spousal caregiver
Back in the 1990s, some psychologists were already raising concerns that many studies weren’t holding up when tested again. At first, these warnings didn’t get much attention, but they planted the seeds of what would later be called the replication crisis.
The issue gained real traction in 2005, when Dr. John Ioannidis published a critique suggesting that a large portion of published research findings could actually be false. His work hit hard, because it wasn’t just pointing out isolated problems — it was questioning the reliability of entire fields of study.
Then in 2011, Dr. Daryl Bem published highly controversial studies with surprising results that many struggled to accept. If research like that could pass peer review, it raised uncomfortable questions about how science was being done.
By 2015, there was no ignoring the problem anymore Prof. Brian Nosek and the Open Science Collaboration launched large-scale replication projects, systematically testing published studies. Their work showed just how widespread the issue really was, making the replication crisis a central topic in modern science.
I first heard about the replication crisis around 2009 through a friend’s father, a neuropsychiatrist. I was struck by how studies in psychology and neuroscience that everyone seemed to rely on were suddenly shown to be fragile, inconsistent, or even completely irreproducible. That experience shaped the way I approach scientific research: never blindly accepting results, but always treating them with a degree of caution.
Researchers often work with small samples and sift through their data until something “significant” appears, while journals tend to favour publishing positive results. This combination means many studies look convincing on paper but cannot actually be replicated.
The infamous dead salmon study showed that neuroimaging can sometimes detect “brain activity” where none exists even in a dead fish being shown pictures of humans expressing emotions. This highlights how small sample sizes, multiple statistical comparisons, and improper controls can produce spurious results. FND research suffers from similar pitfalls. Many studies use tiny participant groups, run numerous analyses, and interpret ambiguous brain activity as meaningful.
Just as the salmon could “light up” in a scanner despite being dead, patients labelled with FND may be subject to conclusions drawn from flimsy or misleading evidence. The dead salmon study exposes the fragility of such methods which helps explain why the evidence base for FND remains unreliable.
For me, this isn’t just an abstract discussion about statistics or replication. My wife’s experience shows how scientific and methodological issues have real-world consequences. When diagnoses are applied based on fragile evidence, they can prematurely close diagnostic doors, leaving patients with uncertainty, misinterpretation, and missed opportunities. Understanding these issues is fundamentally philosophical, calling for keeping diagnostic doors open and maintaining a mindset of careful, evidence-based investigation.