A suite of seminal neuroscience papers by Henri Korn of the Pasteur Institute in Paris allegedly contains a string of anomalies in data interpretation, according to a reanalysis of the papers, published this week in the Journal of Neurophysiology1. But Korn and his co-authors contest this and are critical of the reanalysis, which appears in the same journal as many of Korn's original papers.

A reanalysis of research carried out at the Pasteur Institute casts doubt on a respected hypothesis. Credit: L. BORGHI

The papers, published over the past 25 years by Korn and his co-workers, including Donald Faber of the Albert Einstein College of Medicine in New York, concern the dynamics of the release of neurotransmitter chemicals at the synapse — the junction between nerve cells (see 'Theory of neurotransmitter release moves on'). They suggest that a single bouton (nerve terminus) releases only one quantum of transmitter per nerve impulse. This influential theory has major functional implications, but remains controversial.

A key finding underpinning their theory was based on electrophysiological studies of giant nerve cells in goldfish, called Mauthner cells. Korn and Faber claimed that the number of synaptic boutons counted by light microscopy was highly correlated with the number worked out from an analysis of the amplitudes of the electrical spikes triggered by the neurotransmitter2. But their graph of the correlation, with data points lying on a nearly perfectly straight line, is “almost miraculous” given the noise and uncertainties in the underlying data, claims Jacques Ninio, a bioinformatician at the Ecole Normale Supérieure in Paris, who carried out the reanalysis.

Ninio extracted the data from graphs in the papers and recomputed them. “Several theoretical curves were simply not what Korn and co-workers claimed them to be,” he says.

Ninio's conclusions add to similar allegations by two researchers who worked in Korn's laboratory — Nicole Ropert, now at the University of Paris Descartes, and Luca Turin, a former researcher for the CNRS, France's basic-research agency, now at University College London. In 2004, Ropert submitted a 25-page report to the Pasteur Institute's research integrity committee detailing allegations of “events contrary to scientific ethics”. But the committee last year opted to close the matter without an independent investigation. A similar request made by Turin to the CNRS in 1989 was also not taken forward.

Ninio's challenge is dismissed in an accompanying response3 by Korn, Faber and statistician Alain Mallet of the Pierre and Marie Curie University in Paris, who was a co-author on several of the papers. They criticize Ninio's approach, describing it as “qualitative assessments of second-order representations of the data”.

The 'sheep' mentality is alive and well even at the summits of neuroscience.

“Although we may have made some mistakes — inherent in any scientific inquiry — none of the putative errors invalidates the major findings in our papers,” they write. The research, they write, “changed the nature of the scientific discussion about structure–function correlations at synapses”. They also add: “Subsequent research carried out independently by a number of eminent scientists supported our proposal of the 'One-Vesicle Hypothesis'.”

Nature has obtained the referees' reports on Ninio's paper. One referee comments that Ninio “demonstrates convincingly” that claims in some of Korn and his colleagues' papers are unsuppported. They “are at best erroneous, and at worst deliberate falsifications of the results of the mathematical analysis”, the referee alleges. The second referee's report argues that Ninio raises “a disquieting number of discrepancies” and that Ninio “ventures to say what many experienced observers have politely evaded: that at least one of the emperors of French neuroscience has no clothes”.

Nature put these allegations and referees' comments to Korn, who says he answered the scientific queries raised by Ninio in his published response.

But that rebuttal is “unconvincing, though artful”, claims Paul Adams, a neurobiologist at Stony Brook University in New York. “Ninio did the best he could in view of the fact that he did not have access to the original data.” Adams describes the Ninio paper as “very useful”, saying that published discussions of this issue have not been as sceptical as they should have been. “The 'sheep' mentality is alive and well even at the summits of neuroscience,” he says.