Forensic Examination of Questioned Scientific Images

This paper discusses the Office of Research Integrity's (ORI) experience from its oversight review of institutional scientific misconduct cases involving questioned images, provides illustrative examples of computer image processing techniques with an explanation of how the image analysis contributed to the final misconduct determination in specific cases, and reports a compilation of some generic features of 35 ORI cases (1989-2001). The absolute and the relative incidence of cases with questioned images has increased over the last decade, as has the use of computers for presentation of the questioned image data. To date, ORI has seen no cases of de novo fabrication of images, and fortunately, in most cases the appearance of the questioned image has been quite secondary to the original basis for challenging its authenticity. The survey of cases indicates that good laboratory practices and, notably, attention to mentoring could have minimized the scope of the scientific misconduct when the act of falsification occurred.

Keywords:

• Autoradiograms
• Blots
• Falsification
• Image Processing
• Manipulation
• Scientific Misconduct