Contribution
Before it, functional imaging had shown “islands” of preserved cortical response in some patients diagnosed as vegetative, but a cortical response to a stimulus could be unconscious. This paper replaced the stimulus response with a response to command that has no motor output: the patient was told to perform one of two mental-imagery tasks at specific points in the scan, and the regions each task recruits in healthy people were active when, and only when, she was told to do it. It is a single case, and it proposes, without doing it, that such tasks could be used to communicate.
Key insight
The marker of awareness moved from behaviour to compliance with an instruction, read from the brain. Comprehension alone does not count: the authors set aside their own first result, intact speech and ambiguity-resolution responses, as “suggestive” only, because semantic processing can run without awareness. What counts is that the patient did what she was asked, when she was asked, for a sustained period, with a mental act that has no motor output and so can be produced by someone who cannot move. The decision to cooperate is the evidence.
Assumptions
- Diagnosis. The patient “fulfilled all of the criteria for a diagnosis of vegetative state according to international guidelines” (Royal College of Physicians 2003), by multidisciplinary clinical assessment. The details are in the SOM, not read.
- Region specificity. Tennis imagery reliably recruits the supplementary motor area (SMA), and spatial navigation imagery the parahippocampal gyrus (PPA), posterior parietal cortex (PPC) and lateral premotor cortex (PMC), in healthy people. The paper relies on this and on its own 12 controls.
- Inference from compliance. Sustained, task-specific, cue-locked activity cannot arise without a conscious decision to comply. This is stated, not argued: “her decision to cooperate with the authors by imagining particular tasks when asked to do so represents a clear act of intention, which confirmed beyond any doubt that she was consciously aware of herself and her surroundings” (p. 1402).
- Asymmetry of inference. Only positive findings count: “negative findings in such patients cannot be used as evidence for lack of awareness, because false negative findings in functional neuroimaging studies are common, even in healthy volunteers” (p. 1402).
Key results
- Patient. “In July 2005, a 23-year-old woman sustained a severe traumatic brain injury as a result of a road traffic accident. Five months later, she remained unresponsive with preserved sleep-wake cycles” (p. 1402). Nothing else about her history is in the main text. Monti et al. (2010, NOTE-964) list her as their Patient 4 with an interval since ictus of 6.0 months.
- Speech (first study). Spoken sentences against acoustically matched noise gave speech-specific activity bilaterally in the middle and superior temporal gyri, “equivalent to that observed in healthy volunteers” (fig. S1, not read). Sentences with ambiguous words gave an additional left inferior frontal response, “similar to that observed for normal volunteers”. The authors: this “although suggestive, is not unequivocal evidence that a person is consciously aware” (p. 1402).
- Imagery (second study). Spoken instructions “to perform two mental imagery tasks at specific points during the scan”: imagining “playing a game of tennis” and “visiting all of the rooms of her house, starting from the front door”. Tennis: significant SMA activity. House: PPA, PPC and lateral PMC. “Her neural responses were indistinguishable from those observed in healthy volunteers (fig. S2) performing the same imagery tasks in the scanner” (p. 1402). Controls: a group of 12. Threshold: P < 0.05 corrected for multiple comparisons (Fig. 1 caption; the ”<” is mis-encoded in the extracted text). The number of blocks, their length and the statistic for “indistinguishable” are in the SOM, not read; the main text gives no quantitative comparison with controls.
- No practice. The responses were “reproducible and robust task-dependent responses to command without the need for any practice or training” (p. 1402).
- No communication. No question was put and no answer decoded. The paper proposes that patients “diagnosed as vegetative, minimally conscious, or locked in, may be able to use their residual cognitive capabilities to communicate their thoughts to those around them by modulating their own neural activity” (p. 1402).
Claims
| id | claim | strength | support |
|---|---|---|---|
| C1 | A patient meeting clinical criteria for the vegetative state produced task-specific, cue-locked activity to two imagery commands, matching 12 controls | moderate | one patient, Fig. 1 and fig. S2 (SOM not read); no quantitative comparison in the main text |
| C2 | That response shows she understood spoken commands and responded “through her brain activity, rather than through speech or movement” | moderate | inference from C1 |
| C3 | Her compliance “confirmed beyond any doubt that she was consciously aware of herself and her surroundings” | weak | asserted from C1; no argument that unconscious processing could not produce cue-locked task-specific activity |
| C4 | Intact speech and semantic responses are not by themselves evidence of awareness | moderate | cited literature on implicit learning, priming, anaesthesia and sleep |
| C5 | The method may let some non-communicative patients communicate | weak | proposal; not attempted here |
Concepts
- Vegetative state. Patients “who emerge from coma appear to be awake but show no signs of awareness”; the diagnosis “depends crucially on there being no reproducible evidence of purposeful behavior in response to external stimulation”.
- Act of intention. The paper’s marker of awareness: deciding to do what one is asked when asked, here by a mental act.
Connections
It builds on Schiff et al. (2002) on preserved “islands” of cortical function in vegetative patients and on the Royal College of Physicians’ diagnostic guidance. Monti et al. (LIT-1192, read in NOTE-964) extend it to 54 patients, include this patient as their Patient 4, and build the yes/no procedure the paper proposes. Block (NOTE-497) reports it as “an intentional act without report”. That reading is correct for what the paper shows: there is no report, only an act on command.
Bearing on the record
- CASE-107 and the channel. The response is a mental act with no motor output: she responded “through her brain activity, rather than through speech or movement” (p. 1402). But the act that carries it in the tennis task is motor imagery, and the region it is read from is the supplementary motor area, a motor-planning region. The “spatial” task also recruits lateral premotor cortex. So the evidence of awareness crosses a channel the patient drives, and that channel is imagined movement. This strengthens CLAIM-337’s description of it as action moved inward, in the most literal sense available: the patient was told to imagine acting, and the readout is the motor system’s preparation for an act not made. The thesis’s reply, that it acts on the patient’s own state and not on the world, stands: nothing outside her changed except through the scanner. The CASE’s sentence “the output channel is mental action alone, with no action on the world” is right for this paper.
- What it shows about judgement. It shows comprehension of a spoken instruction, its retention, and its execution at the cued time. It does not show judgement in the record’s sense, an interpretation reached about something and then used: no question was asked, and the patient chose only whether to comply. Owen et al.’s “beyond any doubt” is an assertion, and a reader citing this paper for lucid judgement without action (CLAIM-341) should cite Monti et al. instead, where answers were given.
- The asymmetry. The paper states its own form of it: positive findings confirm awareness, negative ones show nothing (p. 1402). That is the CASE’s asymmetry from the observer’s side: the evidence can only arrive through a channel the patient drives, so a patient who does not drive it is uninformative.
- Prior history. The patient was 23, injured five months earlier; she built her model while acting, as the CASE’s “What it cannot show” says. Five months is far shorter than the years at issue in Kübler and Birbaumer’s extinction hypothesis (LIT-1199), so the case says nothing about whether judgement survives a long loss of contingency between intention and outcome.
- No feedback. The main text describes no feedback to the patient. She complied without any restored contingency between intention and outcome, and “without the need for any practice or training”. Against Chaudhary et al.’s patient (LIT-1200), who communicated only after neurofeedback, this is a response with no contingency supplied in the session. It is not evidence that every contingency was absent before it: the paper says nothing about her bedside behaviour beyond the diagnosis.
- CLAIM-342, QUESTION-121. Nothing directly. The anchor of the patient’s own loop is not addressed.
- The LIT. Its summary is right as far as it goes. It should add that the paper asked no question and decoded no answer, that the comparison with controls is qualitative in the main text, and that the authors set aside the speech result as not evidence of awareness.
- For the anthology. No instruction for machine-learning practice.
Limitations
- One patient. No statistics for the patient–control comparison in the main text; the SOM was not read.
- The inference from compliance to awareness is asserted. The paper does not discuss whether an instruction could trigger task-specific imagery activity without awareness.
- The paper itself says false negatives are common, so it cannot estimate how many vegetative patients are aware.
- It does not attempt communication.
Open questions
- Whether cue-locked, task-specific activity can be produced without awareness. A result showing it in unaware healthy subjects (under anaesthesia or sleep) would remove the paper’s main inference.
- For the record: whether a patient with no motor channel at all, years after injury, can still comply on command. That would bear on CASE-107’s extinction thread.