Lawrence Farwell's method is a Concealed Information Test using electroencephalography rather than sweat glands. He calls it brain fingerprinting, which is an unfortunate name, and much of the trouble that follows comes from it.
The physiology is real. The P300 is a positive-going voltage deflection appearing roughly three hundred milliseconds after a stimulus that is rare and meaningful to the subject. It is among the most studied components in cognitive electrophysiology, it is reliably elicited by recognized items embedded in a stream of unrecognized ones, and it is involuntary. Farwell adds later components and calls the composite a MERMER.
The procedure follows the CIT: probe items relevant to the crime, target items known to the subject, and irrelevant items, presented rapidly while EEG is recorded. A P300 to probes indicates that the information is present in the subject's memory.
That is the correct description of what is measured, and Farwell has generally been accurate about it in technical settings — the method detects a record stored in the brain, not a lie.
The case that made it famous is Terry Harrington's. Harrington had been convicted of the 1977 murder of a retired police captain in Council Bluffs, Iowa, largely on the testimony of a teenager who later recanted. In 2000 and 2001, at a post-conviction hearing, Farwell testified that Harrington's brain showed no recognition of salient crime-scene details and did show recognition of his alibi. Judge Tim O'Grady admitted the testimony while denying relief on other grounds. The Iowa Supreme Court reversed Harrington's conviction in 2003 — on the basis of suppressed police reports pointing to another suspect, explicitly not on the brain evidence — and the charges were dropped.
So the technique has been admitted in one hearing and has never been the basis of a decision.
The independent scientific assessment is more critical than the publicity. J. Peter Rosenfeld, who has worked on P300-based concealed information detection for decades, has published countermeasure studies showing that subjects instructed to make covert mental responses to irrelevant items — imagining a movement, or thinking of a specific memory — can defeat P300 protocols, reducing detection rates substantially. Rosenfeld has also criticized Farwell's claimed error rates and the paucity of independent peer-reviewed replication of the specific proprietary method.
Two lessons come out of this that apply well beyond one instrument.
The first is the danger of a name. Brain fingerprinting invites a comparison to fingerprints, which are individuating physical evidence with an established error literature. The method is not that; it is a probabilistic memory-recognition test with all the operational fragility of chapter 116's CIT, plus countermeasures.
The second is the pattern of the claim. A genuine, well-understood neural signal, an inferential leap from recognition to guilt, a proprietary implementation with claimed near-zero error rates, and marketing well ahead of independent replication. That combination has appeared three times in Part 4 already.
Cross-ref: Concealed Information Test chapter