Reliability And Implications Of Neuroscientific Findings In Legal Contexts
Neuroscience has steadily become a standard part of the psychological assessment in court but is still relatively new. The usage of it still lacks consensus from both the neuroscientific as well as the legal side and proper policies have not been put into practice. The discussion can be divided into two main concerns: how reliable are the conclusions drawn from neuroimaging data and how can these findings be translated into a legal context. Some neuroscientists disagree on what the implications of such usage should be and how much value we should put on our current knowledge. There are concerns on the inaccuracy of the neuroscientific imaging techniques, fMRI being the primary target of conversation, that pose a risk in misinterpreting and naively jumping into causational conclusions with correlational evidence. The fundamental purpose of courts is to establish or debunk the likelihood of someone having performed a certain behavior, i.e. having committed a crime. These questions seek causal relationships that neuroscience cannot (yet) provide. However, such limitations do not mean neuroscience cannot be of any value to legal procedures. It should be continued to be used as a part of standard psychological assessment while being mindful of its nature and limitations. With constantly improving means of investigation, neuroscience is likely to gain more clout in legal contexts. This calls for a strict protocol for the legal implementation of neuroscientific findings to avoid misuse and -interpretation.
A psychological assessment has been a part of determining a defendant’s legal responsibility for a long time. What is much more recent, however, is using findings from cognitive neuroscience (CNS) as a part of that assessment. In an ideal world, cognitive neuroscience would offer an objective tool, increasing the reliability and validity of the psychological assessment. Neuroscience could for example provide new insight on what is the most efficient treatment/punishment verdict for the convict, and additionally, to reduce the risk of malingering, which is a forensic setting can pose a threat to justice (Scarpazza, Pellegrini, Pietrini & Satori, 2018).
In reality, however, bringing neuroscience into the courtrooms is much more complex and a heated topic of debate within the field of law and neuroscience. Some neuroscientists endorse the usage of neuroscience in court and believe it can provide a useful tool (Goodenough and Tucker, 2010; Jun and Yoo, 2018), while others are concerned about misunderstandings or even debunk the idea entirely (Eastman and Campbell, 2007; Gkotsi, Gasser, 2016). In this paper, I am going to discuss the question ”To what extent is the usage of neuroscientific findings justified as part of legal procedures in terms of accuracy of findings?’ I am going to approach this question by first discussing the current limitations of the most used neuroimaging technique we use today, functional magnetic resonance imaging (fMRI), and how these shortcomings can be tackled in the future. Subsequently, I am going to discuss how the context of the law might alter how the results of neuroscientific studies should be interpreted and what possible threats this might pose. Simplifications and over-generalizations of neuroscientific findings pose a dangerous threat to practicing justice.
The main source of imaging evidence used in court comes from functional magnetic resonance imaging (fMRI) (Gkotsi & Gasser, 2016). At the moment, it is widely considered one of, if not the most reliable technique in neuroscience what it comes to investigating the relationship between behavior and physiological changes in the brain (Ekstrom, 2010). That is why it has been proposed that the courts should limit themselves to only use fMRI images (Brown & Murphy, 2010).
Using fMRI is advantageous because it is fairly non-invasive, quite available, and has a relatively high spatiotemporal resolution (Logothetis, 2008). It is also able to illustrate the activated brain areas during a specific task. One shortcoming of fMRI, however, is that it does not measure neural activity directly. It relies on a hemodynamic modality, measuring a surrogate blood oxygen level-dependent (BOLD) signal. This surrogate signal is restricted by two main limitations; accuracy and speed (Ekstrom, 2010; Logothetis, 2008). The fMRI measures mass action rather than individual cells, which makes the results less accurate than desired. The BOLD signal is also slower than neural activity, meaning that it cannot measure neural activation in real-time (Logothetis, 2008). An additional shortcoming of using the BOLD signal is that it might be confusing activation with inhibition of neurons (Brown & Murphy, 2010). Although inhibition is brain activity in the sense that the brain is actively inhibiting the concerning neurons, confusing the two might have important implications in analyzing the data (Murphy & Brown).
In just a few years, however, there have been major improvements in fMRI due to the work with 7 T. Compared to the common measuring at 3 T and 4 T, measuring at 7 T can provide a better isotropic resolution of 0.5. Additionally, it improves the signal-to-ratio (SNR), thus reduces noise from surrounding areas (mass action). This consequently results in improved spatial precision in the BOLD signal (Hutton et al., 2011; Turner, 2016). Combined with complex techniques such as parallel acquisition and spin-echo (SE), fMRI can now produce images of even deepest structures with high spatial resolution and impressive speed (Turner, 2016). Although a substantial proportion of highly cited neuroscientific studies were conducted at 3 T or even at 1.5 (Logothetis, 2008), scanning at 7 T is being more and more implemented and increasing the quality of results tremendously (Turner, 2016). Logothetis (2008) summarizes, that fMRI is a magnificent method, but can only be truly reliable in exploring precise mechanisms when combined with other techniques, such as animal research.
Alongside combining fMRI with other means of investigating, it attempts to overcome some of its limitations by meta-analysis and averaging results. It is important to understand that at the moment, neuroscientific evidence from fMRI can offer correlates of what is the likelihood inside a population, not causal relationships between a brain area and behavior in an individual (Buckholtz & Faigman, 2014). Eastman and Campbell (2007) argued that this can be problematic in the case of law, which in contrast seeks causal links for the behavior of an individual rather than that of a population. They proposed that because neuroscience and the court of law are fundamentally asking different questions, they cannot be combined in a non-biased way, a problem which they referred to as ‘a a mismatch’.
How well law and cognitive neuroscience can work together has been a matter of debate between scientists. Eastman and Campbell (2007) argued in their paper that there is a fundamental mismatch between what questions the court of law is asking and what kind of questions CNS is answering. The court deals with individual cases and tries to determine what kind of a person the defendant is and to what extent they are legally responsible for the crime they have potentially committed. Cognitive neuroscience, on the other hand, deals with populations and associations; findings from studies of this nature cannot answer questions about an individual case. As an example, if neuroscientific evidence shows that having abnormalities in brain area X is linked to increased aggression, the court could conclude that if the defendant shows abnormalities in brain area X, their violent behavior was partially influenced by this fact. While this may be true, this should not be concluded from neuroscientific findings, which cannot say anything definite about individual behavior (Brown & Murphy, 2010). Eastman and Campbell (2007) argue that ‘translation from science to law’ can easily go askew and they question the use of such findings as substantial evidence. When analyzing a brain scan that shows abnormalities, one cannot say that these abnormalities caused the individual to exhibit the behavior in question, it merely increases the likelihood. Further, when translating neuroscientific findings into a criminal case, the questions shift from ‘likelihood in a population’ into ‘likelihood for this individual’, which from a scientific point of view is problematic and not reliable. Using brain scans that are made after the incident also poses another issue. Brown and Murphy (2010) argue, that using fMRI images to explain the defendant’s mental states at the time of the crime is problematic, as science seeks objective answers of reality, not interpretations. They argue that although fMRI might be a valid tool in medicine and research, it is not reliable enough to explain the complex cognitive states of an individual, especially when done so in retrospect.
Goodenough and Tucker (2010) propose a contrasting argument, saying that law and cognitive neuroscience are ‘natural partners’; the law constantly seeks the help of other fields; it is its fundamental purpose to get the expert option from many angles and in this way practice justice as well as possible. Psychology has also been involved for quite some time, and Goodnenough and Tucker (2010) foresee that how well neuroscience and law can work together depends on how well the fields can collaborate. They call for projects where experts from both fields and even some from fields such as philosophy combine their knowledge. Many neuroscientists have voiced their concern about the misinterpretation of neuroscientific evidence when used outside the scientific contexts and that is where a multidisciplinary view could help with.
In a quite recent paper, Jun and Yoo (2018) propose a similar stance and offer techniques to make better use of fMRI data, which also could help to integrate neuroscientific evidence into courts. They argue that although neuroscience cannot yet provide definite, causal answers, it does share the same goal with the court of law: they both seek to certify the emblematic reliability of neuroscientific evidence. Jun and Yoo (2018) introduce three strategies neuroscience uses to explain phenomena: cognitive subtraction, data-driven approach, and brain manipulation approach. They predict that new brain manipulation approaches will be implemented e.g. optogenetics when limitations of accessibility, invasiveness, and quality will be tackled. They further argue that cognitive subtraction, i.e. manipulating some cognitive function and observing a brain area, is outdated due to only controlling for the consequence; the cognitive process. However, data from this technique can be useful when taking a data-driven approach.
The mass-data approach employs extensive databases and with refined, computerized data-processing can increase the reliability of relationships between brain and function. Meta-analyses were done by manual work only go through a limited amount of data. Computer-processing that uses artificial intelligence and machine learning is not confined by the physical constraints of a human, thus being able to process and analyze data dar more efficiently (Jun & Yoo, 2018). There are already many data platforms that are automatically analyzing neuroimaging evidence, e.g. OpenfMRI (Poldrack, 2011) and NeuroVault (Gorgolewski et al., 2016). Some ongoing projects are attempting to shift from area-specific research into including the whole brain (Del Pinal and Nathan, 2016).
They further point out that meanwhile the technology is not quite there yet to provide causal relationships, it is not necessarily what the court needs to count something as evidence. It is stated in the U.S. code of law in the federal rule of evidence 401, the test for Relevant Evidence (1975) suggests that even the neuroscientific evidence, which only points out to correlates in the population could be applied: ”evidence is relevant if it has any tendency to make a fact more or less probable than it would be without the evidence”.
Techniques used in neuroscience are constantly improving; perhaps someday earn a spot next to DNA as substantial evidence. Meanwhile, ‘more probable than without the evidence is still better than nothing. Neuroscience has limitations that we have to be mindful of, but when used alongside traditional psychological evidence and other relevant evidence for that specific case, it does have value. If we do this, neuroscience can help us to reach better justice.
Interpreting neuroscientific findings can easily lead to bias due to the limitations of our current technology. Using such evidence in courts poses an even higher risk, given the fundamental nature of the legal procedure; legal cases deal with individual cases and are inclined to draw causal relationships between matters. For these questions, neuroscience cannot yet provide direct answers, but that should not be a reason for not using such evidence at all. The voxel resolution and speed of fMRI are constantly improving as well as the computerized methods to analyze data more efficiently. The computerized data-analysis methods also make it possible to analyze the brain as a whole thus tackle the bias of area-specificity. This will also improve the reliability of neuroscientific evidence in legal contexts and most likely increase the influence it will have on verdicts. Meanwhile, we have to be mindful of the limitations of fMRI imaging when we implement neuroscientific evidence in courts. Neuroscientific evidence should be used alongside other assessments and direct causations should not be drawn between neuroscientific images and the behavior of the defendant.
We should not abandon the usage of neuroscience as part of the court of law merely because the issue poses hard questions. However, we have to acknowledge we live in a transformative stage, where the accuracy of neuroscience cannot (yet) provide impeccable answers and the etiquette on how to integrate neuroscience into courtrooms is not well established. Clarification on the usage of neuroscientific findings must be achieved to defeat the current obstacles. To defeat these obstacles, further research is needed to make brain imaging techniques more reliable and to gain a better understanding of how neural activity relates to behavior. Courts call for evidence that can analyze the behavior-neural activity relationship on an individual level. Further, a systematic procedure on how to include and use NS in a legal setting should be inaugurated. Meanwhile, our technology is advancing, we should be mindful of the current limitations and treat neuroscientific findings not as absolute truths about the individual, but as correlates of the group the suspect is a part of.
Cite this Essay
To export a reference to this article please select a referencing style below