Andrew Peterson and Tim BayneHowever, the translation of these methods to the clinical setting raises a number of chal-
lenging epistemic questions. For example, decisions will need to be made about how much
evidential weight should be given to covert command-following, the capacity to follow nar-
rative structure, or to detect violations of global regularity when it comes to the ascription of
consciousness.
These questions are challenging, but they are not different in kind from those that we face in
determining how much weight to place on behavioral measures of consciousness. (Consider, for
example, the debate regarding visual fixation and pursuit as markers of consciousness, which we
noted in Section 3.) Indeed, in certain cases—for example, with respect to covert command-
following and overt command-following—these questions are essentially variants of a single
question about clinical validation: how do we know that any given assay of consciousness is
actually measuring consciousness?
In previous work, we have articulated what we have variously called the “consilience”
(Peterson 2016) or “natural kind” approach (Shea and Bayne 2010) to questions of validation.
In essence, this approach holds that novel measures of consciousness should be validated on
the basis of their fit with other putative measures of consciousness. Greater evidential weight
should be given to novel measures that produce evidence that is conciliate with other putative
measures, while less evidential weight should be accorded to novel measures that are less strongly
associated with other putative measures of consciousness. A certain kind of circularity is inher-
ent in this approach, but this is arguably a form of circularity that is inherent in any attempt
to validate novel measurement techniques in domains in which we lack “gold standard” tests.
Recently, several research teams have adopted this approach in their own neuroimaging and
EEG studies (Sergent et al. 2017; Chennu et al. 2017; Demertzi et al. 2017).
Another way in which neuroimaging and EEG research promises to reshape how we think
about the recovery of consciousness following brain injury concerns the very diagnostic cat-
egories that are used in neurology (Bayne et al. 2017; Peterson and Bayne 2017). As we have
seen, neurologists recognize only two categories in this area: the VS and the MCS. It is now
clear that this taxonomic system is overly simple, and that the field would benefit from the
development of a more nuanced taxonomy that reflects the spectrum of conscious states and
capacities that can be found in post-comatose patients. Some patients exhibit low-level percep-
tually guided behaviors, such as visual pursuit and fixation. Some patients exhibit command-
following behaviors. Some patients appear to be able to follow the plot of a film. Indeed, even
further distinctions need to be made within these categories, for some patients have the capacity
to engage in overt command-following but lack the capacity to engage in covert command-
following, whereas other patients manifest precisely the opposite suite of capacities (Gibson
et al. 2014; Monti et al. 2010). Further, the fact that a patient is sensitive to the narrative struc-
ture of a film does not imply that she can communicate through neuroimaging (Naci et al.
2014, 2016). Although some of the capacities that are probed by these paradigms are hierarchi-
cally related, others are not.
An adequate taxonomy for post-comatose disorders of consciousness should capture these
variations in cognitive capacities. The recent introduction by Bruno et al. (2011b) between
two types of MCS patients, MCS+ patients and MCS− patients, is a first step in this direc-
tion. Bruno et al. (2011b) suggest that the label “MCS+” should be reserved for patients who
can follow commands, whereas the label “MCS−” should be reserved for patients who lack
this capacity but show other behavioral signs of consciousness, such as visual fixation and
pursuit or the ability to localize noxious stimuli. Although these categories have not yet been
incorporated into neurological taxonomy, in our view there is every reason for this step to
be taken.