Measuring consciousness

I have been reminded about the business of measuring consciousness with EEG machines discussed at reference 2, by the short article – which seems to have appeared in various magazines, with my copy coming from some part of the ‘Nature’ empire – at reference 1.

One of the various machines that can do this is described at reference 5, by the people (Medtronic) selling the machine in question. A machine which calculates an index of consciousness, called the bispectral index, from EEG signals. A machine which is intended to be used by anaesthetists during operations to help them maintain the right level of anaesthesia. A machine which I noticed in passing in the posts at references 3 and 4, back in 2016.

The machine described at reference 2 claims to do rather more than this bispectral index, ultimately derived from work done by physicists half a century ago now on problems like forecasting the weather and analysing seismic signals. Claims that this machine can distinguish, for example, between those in a truly vegetative condition and those who are merely locked in - clearly an important matter from a clinical point of view, from the point of view of determining the treatment appropriate for a patient in one of these conditions.

Rather more also in the sense that it can so distinguish without needing to put the patient in a scanner, in the way reported at references 9 and 10, it being enough to put the patient in the far less cumbersome EEG cap, not that different from the sort of cap that some ladies wear when swimming.

On the other hand, it does not look as if it had then (reference 2 being five years old now) been worked up to real time application (as in during an operation) and it does look to be a lot more invasive, involving poking the brain with pulses of electricity. Not passive in the way of regular EEG at all – although I guess we should remember that scanners are not very passive either, involving lots of high powered pulses of something or other.

All that apart, the tricky bit for me in the present paper was the computation of something called the SS matrix, a two dimensional description of the response of the brain to its poking, in space (across the cortical sheet if not through the brain) and in time. With ‘SS’ standing for significant source and with these significant sources (places in the brain, or on the cortical sheet) having been extracted from the 60 channels of EEG recording by means of a good deal of statistical trickery. Having got the matrix you then compute its complexity and there you are. The more complexity you have, the more consciousness you have – with this result being worked up and validated with around 50 subjects in a range of states of health and consciousness. With all this being summarised in the four boxes of the diagram above.

I am not here concerned with the substance of this paper (that is to say, reference 2), rather with the implications for LWS-N (reference 7) of a conclusion that consciousness requires complex activity right across the brain. This because LWS-N hypothesises consciousness arising in one fairly small part of the brain, very probably, if only for considerations of symmetry, somewhere in the middle of the brain, in one of the older structures in or near the top of the brain stem.

LWS-N also hypothesises that one can separate the building of the data content of consciousness from that content’s projection as a subjective experience, and LWS-N is about the organisation of a compact data structure, expressed in neurons, the activation of which amounts to consciousness, with this activation being the projection in question. It is not about all the work needed to build that data structure, work which will often include, for example, all the work that is done between the retinas and the visual parts of the brain. Work which interacts with all the other work going on at the same time and with memory. We refer to all this, after the fashion of the 1970’s, as compilation, with it being the results of that compilation that finds its way into consciousness. So it certainly is the case that this compilation process does require complex activity across the brain.

Furthermore, the activity measured in the course of the work reported at reference 2 is the direct result of poking the brain, stimulating it electrically. So we are not saying that this complex activity is consciousness, rather that the state or condition of consciousness requires that all the connections needed for stimulation of this sort to reverberate around the brain should be in place and ready to go.

So maybe there are no direct implications for LWS-N. Although one does wonder why the brain bothers with concentrating a very small proportion of the data available to it in this rather peculiar way. Or put another way, what is consciousness for?

Koch also observes in the course of reference 1 that: ‘… IIT also predicts that a sophisticated simulation of a human brain running on a digital computer cannot be conscious – even if it can speak in a manner indistinguishable from a human being. Just as simulating the massive gravitational attraction of a black hole does not actually deform spacetime around the computer implementing the astrophysical code, programming for consciousness will never create a conscious computer. Consciousness cannot be computed: it must be built into the structure of the system…’. Where IIT is a theory about consciousness originally put forward by the same Tononi who figures in the list of authors of reference 2 and who was first noticed by me, back in 2013, at reference 8. And properly described at reference 12. I think it is fair to say that LWS-N makes such a prediction too, with its hypothesis being that consciousness arises from some peculiar electrical activity in a small patch of cortex, most unlikely to be replicated by any sort of computer presently in view. So robots are not going to be conscious any time soon, although it does seems likely that they will soon be able to behave as if they were.

PS: along the way I have been entertained, if not much educated, by knocking up some VB code in Excel to do a simple version of the Limpel-Zev complexity of reference 2. A reminder, if nothing else, of the power of a tool like Excel, a tool which one uses enough in the ordinary course of events that one can still use it to knock up a bit of half decent code. Also of the limits of the Internet, with diligent search failing to turn up a copy of reference 6 for less than $30 or so, rather too rich for me. And with the corresponding entry in Wikipedia largely resisting my comprehension.

References

Reference 1: What is consciousness - Christof Koch – 2018.

Reference 2: A Theoretically Based Index of Consciousness Independent of Sensory Processing and Behavior - Adenauer G. Casali, Olivia Gosseries, Mario Rosanova, Mélanie Boly, Simone Sarasso, Karina R. Casali, Silvia Casarotto, Marie-Aurélie Bruno, Steven Laureys, Giulio Tononi, Marcello Massimini – 2013. Are they so democratic as to give their laboratory technicians and engineers a mention?

Reference 3: http://psmv3.blogspot.com/2016/11/a-day-in-life-of-brain.html.

Reference 4: http://psmv3.blogspot.com/2016/11/an-assembly-for-consciousness.html.

Reference 5: http://www.medtronic.com/covidien/en-us/clinical-education/catalog/monitoring-consciousness-using-bispectral-index-during-anesthesia-bis-pocket-guide-clinicians.html.

Reference 6: On the complexity of finite sequences - A. Lempel and J. Ziv – 1976.

Reference 7: http://psmv3.blogspot.com/2018/01/an-introduction-to-lws-n.html.

Reference 8: http://psmv2.blogspot.com/2013/04/phi.html. I think this remains a fair – if short – summary of my position on IIT.

Reference 9: http://psmv3.blogspot.com/2017/09/ruminations.html.

Reference 10: Detecting awareness after severe brain injury - Davinia Fernández-Espejo and Adrian M. Owen – 2013.

Reference 11: An introduction to bispectral analysis for the electroencephalogram - Jeffrey C. Sigl, Nassib G. Chamoun - 1994. Another failure to find a freebie.

Reference 12: Consciousness: here, there and everywhere - Tononi & Koch – 2015. Royal Society, Philisophical Transactions B, a very respectable journal. Where it is explained that the five axioms of IIT are about intrinsic existence, composition, information, integration and exclusion. With much talk of consciousness being both differentiated and integrated, talk from which I associate to the Holy Trinity, the three which are both three and one. Perhaps the Greeks who invented it were onto something after all.