Cherry-Picking Memories: Why Neuroimaging-Based Lie Detection Requires a New Framework for the Admissibility of Scientific Evidence under FRE 702 and Daubert

Article excerpt

ABSTRACT

Neuroimaging techniques have been in heavy rotation in the news lately. Increasingly, companies have used neuroimaging techniques--specifically, functional magnetic resonance imaging (fMIRI)--in an attempt to determine whether an individual is telling a falsehood. More troublingly, these companies have proffered factual conclusions for use injury trials. This Article discusses the capabilities and limitations of the technique. In doing so, the Article also discusses why the technology will require the federal judiciary to reevaluate its current interpretation of Federal Rule of Evidence 702 and the Daubert doctrine for admitting novel sources of scientific evidence.

TABLE OF CONTENTS

INTRODUCTION

I:   TECHNICAL AND SCIENTIFIC BACKGROUND
     A. Technical Background
        1. Fundamentals of fMRI
        2. Mechanics and Experimental Methodology of fMRI-Based Studies
     B. Scientific Background

II:  NEUROIMAGING EVIDENCE OF LIE DETECTION IN THE FEDERAL JUDICIARY
     A. Legal Admissibility of Neuroimaging Evidence is Mixed
        1. Executive Background
        2. Judicial Background
     B. Functional MRI-based Lie Detection is Neither Reliable Nor Valid
        1. Technical Concerns
        2. Scientific Concerns
        3. Epistemological Concerns
        4. Practical Concerns
     C. Lie Detection via Functional Neuroimaging is Uncertain Under
           Daubert
        1. The Technique Should Have a Clearly Defined and Low Error
           Rate
        2. The Technique Should Have Standards Controlling Its
           Operation
        3. The Technique Should Be Testable or Falsifiable
        4. The Technique Should Have Survived Peer Review and Be
           Accepted Within the Relevant Field
        5. Limitations of FRE 702 and Daubert's Four Factors
     D. Improving Daubert: A New Model for Scientific Validity Under
        FRE 702
     E. Lie Detection via Functional Neuroimaging is Not
        Admissible Under the Proposed Model

III: RECOMMENDATIONS AND THE WAY FORWARD
     A. Why Not Just Let it in For What it is Worth?
        1. Probative Value is Outweighed by Prejudicial Nature
        2. Judicial Efficiency is Not Promoted
        3. Differences of Degree, Not Kind
     B. New Scientific Methods Require a Restrained Approach

INTRODUCTION

It is the dream (or nightmare) of every trial lawyer. A witness is placed into a black box; a question asked ("Did you kill Mr. Smith?"); a response given ("No, I did not"); and then a klaxon blares, dissonant enough to rouse the most catatonic juror, accompanied by an unavoidable flashing red sign: "THAT'S A LIE."

How much faster would trials be, how much less costly the proceedings, how much more justice done, if only witnesses always told the truth? Or, the next best thing: If they could be tested by a lie detector with perfect accuracy and reliability? But how does the story sour if the witness could deceive the machine by pressing his toe onto a thumbtack placed in his shoe? Or the expert administering the test could, like a carnival operator, place a foot on the guy-wire, weighing the answer toward one side or the other?

Two private firms trying to make the dream of perfect truth verification into a reality have recently proposed that functional neuroimaging, a technology used in medicine and cognitive neuroscience, can be used to distinguish truth from deception. (1) Functional neuroimaging records brain activity in a specific location across moments in time. These firms offer to detect deception (or "verify truth") by matching an individual's own pattern of brain activity during interrogation to generalized patterns of brain activity observed when people are known to be engaging in deception. (2) Because functional magnetic resonance imaging (fMRI) is generally considered the most popular functional neuroimaging technique, both firms have adopted its use. (3)

Part I of this Article provides an overview of the technical and scientific background relevant to fMRI-based lie detection in order to apply it to existing doctrinal standards. …