Electrical stimulation of the human cortex: The farther removed from sensory input or motor output structures, the less likely it is that a region contributes to consciousness

Hot or not. Christof Koch. Nature Human Behaviour, Jul 2020. https://www.nature.com/articles/s41562-020-0925-7

Abstract: Electrical stimulation of the human cortex, undertaken for brain surgery, triggers percepts and feelings. A new study documents an ordering principle to these effects: the farther removed from sensory input or motor output structures, the less likely it is that a region contributes to consciousness.

The elicitation rate, the fraction of electrodes the participant can sense when activated (within the safety limits of the stimulation protocol), varies across the accessible cortical surface. It is as high as two out of three electrodes above visual and somatosensory areas and as low as one out of five electrodes above limbic areas or one out of six over the anterior prefrontal regions. When electrode locations were projected onto a 7- or 17-region parcellation of cerebral cortex (derived from functional MRI resting state functional connectivity of a thousand healthy adults), a compelling pattern emerged: the elicitation rate decreased monotonically along a functional-anatomical gradient, starting with sensory regions at the bottom and ending with transmodal, default-mode and limbic networks at the top. The higher up in the cortex a region is, the less likely gentle brain stimulation there will be noticed by the participating brain (Fig. 1). The silence of these frontoparietal, limbic and default mode networks is remarkable as they are thought to be central to much of cognition.

Furthermore, the varieties of distinct experiences (for example, a visual phosphene, a recall of a song, a feeling of unease) increased when ascending this gradient: while the majority of evoked responses in sensorimotor areas reflect the appropriate visual, somatosensory or motor modality, the smaller number of experiences evoked in limbic, midline and the farthest forward prefrontal region were the most diverse across participants. A variety of controls, such as sham trials and varying the amplitude of the iES, ruled out systematic confounds, such as participants having different rates of false alarm or excitability of the underlying tissue varying systematically with location.

While iES is safe and effective, it is also crude: the electrodes are many square millimetres in area and deliver up to 10 mA of bipolar current between adjacent electrodes that can modulate the excitability of a million or more pyramidal neurons and interneurons within a volume given by the resistive spread of the current, supplemented by more remote effects caused by evoking spikes in axons of passage. Still, effects induced by iES can be quite localized, with responsiveness changing from all to none within millimetres or across a sulcus3,7. The challenge for the future will be to move towards microstimulation, common in laboratory animals, in which a thousand-fold-smaller current is sent through thousand-fold-smaller electrodes to give rise to ever more specific sensations. Perhaps this will reveal the remarkable absence of auditory percepts when stimulating Heschl’s gyri, in the neighbourhood of auditory cortex.

The exacting data collected by Fox and colleagues provides critical causal, not just observational, evidence to identify the neuronal correlates of consciousness. Indeed, whether or not the epicentre of experience is in a postulated posterior hot zone or in prefrontal cortex8,9 can be addressed in this manner.