On June 12, 24-year old Aditi Sharma and her boyfriend Pravin Khandelwal were convicted of the murder of Sharma’s ex-fiancé Udit Bharati. Their trials marked the first time brain scans presented as evidence of a defendant’s knowledge of a crime led to a conviction. The lie detection test, the Brain Electrical Oscillations Signature test, or BEOS, was created by Champadi Raman Mukundan, the former director of the psychology department at the National Institute of Mental Health and Neuro Sciences (NIMHANS) in Bangalore, India. The Indian media praised the use of this technology—the Times of India called BEOS a “forensic tool” the “state police can now bank on…to achieve speedy convictions” and declared “the findings clearly indicated their [Sharma and Bharati’s] involvement in the murders”—but neuroscientists in India and elsewhere were wary.Unlike traditional polygraph tests, which detect lies through indirect means such as heart rate and skin conduction measurements, the BEOS test is meant to directly show brain activity. A suspect is read a series of statements that should only trigger memories in someone who actually took part in the events. For example, Sharma was read the statement “I bought arsenic.” Mukundan claims that experiential memories triggered by such statements can be visualized on brain scans and provide evidence of guilt, but a six-member committee led by the current head of NIMHANS called BEOS “unscientific". Other evidence against Sharma and Khandelwal was used, such as polygraph tests, but they insist they are innocent. Mukundan claims that the BEOS error rate is only 5% but his work has not been independently verified and has never been published in a peer-reviewed journal. Few specifics about his methodology are available and it remains unclear how he determined that error rating and if it is reliable for all suspects. His test depends on the detection of experiential memories, and since children often cannot accurately recall how they received certain information (whether they learned it themselves through experience or whether they were told it) it seems plausible that this test will not be accurate for people with certain mental defects. Exactly how sensitive the test is also needs to be explored. Could someone who has repeatedly imagined killing someone but not actually done it be falsely accused on the basis of this test? What if someone has dreamed of committing a crime? However, the most disturbing thing about the use of this test is not that it might not be 100% accurate- it is that judges who allowed the tests to be used ignored the consensus of experts and used evidence that as nonscientists, they cannot fully understand.