The field of functional neuroimaging has made tremendous progress, and contributed a great deal toward furthering the brain sciences. It has become a very useful and valuable tool. Yet, it has a long way to go to reach its full potential. The fundamental hurdle to overcome, I argue, is the inability to connect the mind to the brain. The nature of the relationship between the mind and the brain is far from understood (Bassett & Gazzaniga, 2011). Therefore, neuroimaging is far from creating accurate neural signatures or correlates of human perception, emotion, executive control, goals, attention, imagination, and the rest of the mind.
It is somewhat puzzling that the mind as expressed in real time is not able to be pinpointed in the brain. In theory at least, this is possible. After all, the idea mental states and functions are reflected by a given brain activity i.e. a neural signature is basic cognitive neuroscience.
However the problem of accurately connecting mind to brain — either encoding mind to brain, or decoding brain activity to mental functions — cannot be solved via task-based neuroimaging. There is great difficulty in isolating specific mental functions using psychological tasks (Poldrack & Yarkoni, 2016). A particular task CANNOT be used to define or infer a subject’s mental state.
Functional neuroimaging is by and large task-based, or resting state. However, toward solving the problems above I argue there is a viable third option: mind-based neuroimaging. The subject’s state of mind I argue is a far superior target for neuroimaging than task the mind is engaged in, because the brain signal reflects mind (much) more than the task. A task occurs within a larger state of mind context: perceptual, cognitive, emotional, etc. Task comprises only one part of the mind. Therefore, the mind more strongly and directly correlates with the brain signal than the task.
Overall, the brain signal is a much stronger reflection of the mind — perception, meaning, emotion, motivation, level of focus, fatigue, pain/pleasure, intention etc. — than the task in which the mind is (only partially) engaged.
Let’s now consider the potential use of mind-based functional neuroimaging, through the lens of brain computer interface (BCI) technology. BCI heavily relies on brain signal labeling and classifiers. A BCI device translates the brain signal of the user into external action. The signal — obtained via neuroimaging such as EEG — is analyzed, processed, decoded, and classified. If a classifier matches the signal, this match triggers a corresponding external action, such as an artificial limb movement.
What creates the user’s brain signal? I argue it is the mind as a whole. Of course, the brain signal is influenced by task. But the brain’s (large scale) neural activity is also affected by the environment, measurement situation (social, work, home, lab), recent success/failure at BCI control, and other context. These external variables then affect the mind — the user’s degree of focus or concentration, frustration, fatigue, reward, confidence, relaxation etc. In fact a number of external variables combine with internal (mind) variables, to paint an overall state of mind canvass. These variables and their manifestation in the user’s mind simultaneously affect his or her brain signal.
It is the mind overall, not just the task, that needs to first be defined accurately, before a user’s brain signal “signature” can be defined. Accurately defining the mind and mapping it to the brain requires a new approach. Cognitive neuroscience, despite the best efforts of highly-skilled and knowledgeable experts, is very limiting when applied to BCI development and use. The mind is considerably compressed, into unnatural categories. Working within the current conceptual framework yields only vague and inaccurate definitions of mind. Classifying or labeling the resulting brain signal suffers accordingly. This is no one’s fault. The vast knowledge of the cognitive neuroscience community regarding mind and brain are highly valuable. The problem is this knowledge is being understood and used within the wrong paradigm.
The lack of a clear and accurate mind/brain connection is manifest in most of applied neuroscience, including BCIs. For example, current brain signal classifiers for BCI control are either ill-defined, or too task-focused. A classifier might be “move the cursor to the left” intention (not a bad start, but very incomplete), or “attend to, and do, whatever you did before to make that action happen” (vague and ill-defined). Or, the target might be 100% task-based, such as “whatever the brain does when the user performs a move cursor left task successfully, under conditions Y and Z.”
Brain signal labeling based on the user’s mental state is even more accurate and precise when based on a set of general memories. For instance, the intention “(me) move the cursor left” will include the memories “cursor,” “a computer screen,” and “move left.” This intention — a general memory in its own right – is comprised of a set of general memories. These are mainly visual and somatosensory. Among these are “seeing an object move left,” and “the feeling of my right hand & mouse as it moves left.”
The above memories have strong associations. These include visuosomatosensory attention (“focus on the cursor”), visual imaging (“imagine it will move left”), and emotion (ex: “motivated and focused,” “frustrated,” or “calm and confident”). These associated mental states are also general memories.
Labeling or classifying the brain signal based on task has been successful to some degree. But why settle for this, when it can be more accurately labeled based on the mind? Moreover, I argue it’s possible to define a memory set — its components and associations — for any (common) movement & its context. Movement context would include the person, lifestyle, recent life history, situation (environmental, social, work/home/lab…), and stage of learning (beginner/intermediate/advanced).
Overall, neuroimaging based on the mind, not the task, has the potential to greatly enhance applied neuroscience, including specific BCI projects. Mind-based neuroimaging is, I argue, a potentially powerful tool for BCI development and performance — both in the lab and the real world.
Bassett, D.S., Gazzania, M.S., (2011). Understanding complexity in the human brain. Trends in Cognitive Sciences,15 200. https://people.psych.ucsb.edu/gazzaniga/PDF/Understanding%20the%20Complexities%20in%20the%20Human%20Brain.pdf
Poldrack, R.A., Yarkoni, T. (2016). From brain maps to cognitive ontologies: informatics and the search for mental structure. Annual Review of Psychology, 67, 587. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4701616/