It has been proposed that consciousness is mediated at the level of the neocortex according to a string of declarative, conscious units, which encodes a sequence of sounds or visual objects. A single electric pulse delivered to a neocortical pyramidal fibre after a brief discharge of action potentials renders the fibre inactive for 100 ms or so, since the pulse activates a collateral that engages GABAergic neurons that inhibit the fibre for the purpose of excitability regulation (Chung and Ferster 1998; Krnjević 1974; Krnjević et al. 1966abc; Krnjević and Schwartz 1967; Schiller and Malpeli 1977; Tehovnik and Slocum 2007a). If pulses are delivered in a 10-Hz train, then an activated pyramidal fibre can be inhibited for up to 100 ms between pulses (Logothetis et al. 2010). Thus, 10-Hz stimulation can be used to inhibit a declarative, conscious unit as an animal (including a human) is made to execute a task that is based on a sequence of events such as concatenated sounds or a movie clip, each perceived, imagined, or evidenced using the motor system.

The neocortex is composed of vertically-aligned pyramidal fibres (20 to 40 neurons at any one depth) that are grouped in micro-columns, with each column measuring 30 μm in diameter and believed to encode a single feature (Peters and Sethares 1991). If one uses indwelling electrodes to pass current, a small region of the neocortex (i.e., a 100 to 200 μm diameter sphere of tissue) can be activated with currents of 2 to 5 μA (@ 0.2-ms pulse duration) to evoke or disable perception, which is estimated to drive from 60 to 250 vertically-aligned pyramidal fibres (Peters and Sethares 1991; Schmidt et al. 1996; Tehovnik and Slocum 2007ab, 2013).[1]

As discussed in a previous chapter, the electrical stimulation work of Penfield and Ojemann (Ojemann 1983, 1991; Penfield and Roberts 1966) has been central to the idea that elements in the neocortex can be inhibited by activating specific loci of the cortex to interrupt the generation of speech in alert patients.[2] One of the most important observations made in these studies is that information pertaining to language is stored uniquely in the neocortex: no two individuals have the same language map. This makes good sense, since learning a language (or learning any other faculty) is based on the history of learning, as well as genetic makeup. Therefore, to deduce what is stored within the neocortex of an individual, different declarative, conscious units must be interrupted electrically in the neocortex for different streams of consciousness. This will be no easy task, since the neocortex of humans has a storage capacity of tens of trillions of bits.[3]

To make this line of work more manageable, an understanding of how language is stored in the neocortex is paramount. Each faculty whether it is a distinct language or mathematics is stored by an independent network of neurons (Ojemann 1983, 1991; Rojas 2021). And based on how a language is learned from childhood one can spontaneously develop a verbal network without developing a reading and written network.[4] This suggests that every faculty is anchored to specific sensorimotor transformations: speaking is dependent on sound and the vocal apparatus, reading is dependent on vision (and audio for some) and eye movements, and writing is dependent on vision and hand movements (see: Ojemann 1991).

So, how does the foregoing generalize to other species? Elephants, dolphins, and whales have an advanced communication system that has yet to be deciphered (e.g., Antunes et al. 2011). The songbird, however, has a well-studied telencephalon that is known to store songs (Goldman and Nottebohm 1983; Rochefort et al. 2007), whereby the methods of Penfield and Ojemann could be used to interrupt various aspects of song generation. Thus, the stream of consciousness can be studied across different species by disabling declarative, conscious units electrically at various locations along the neural strings per species, but for this to be feasible the ethology of an animal must be well understood as it is for humans [5] and songbirds.

Finally, since information is ultimately stored at the synaptic level (Hebb 1949), methods will need to be developed to disable a single synapse to study consciousness; it is the synaptic connectivity of a neuron that determines the context within which a declarative attribute is defined per neuron and each human neocortical neuron has, on average, about 10,000 synapses.

Footnotes:

[1] Distinct colors can be evoked from the visual cortex using currents between 2 and 10 μA (Schmidt et al. 1996). To readily evoke such perception using low currents, high-impedance electrode (i.e., > 2 MΩ) that induce a high charge density are recommended (Tehovnik et al. 2009).

[2] Typically, electrical stimulation delivered to the neocortex was used to identify the language areas of the neocortex in patients who were about to have regions of the neocortex removed to treat severe epilepsy (Ojemann 1983, 1991; Penfield and Roberts 1966). In these studies, naming, reading, verbal memory, mimicry of facial movements, and phoneme identification were assessed per stimulation site, typically in Wernicke’s and Broca’s areas. The map size for the primary language was always smaller than the map size for the secondary languages (this difference is related to automaticity as argued in an earlier chapter). Surface stimulation was used with large-tipped electrodes; therefore, the current was in the milliampere range (1-4 mA), and the duration of stimulation train (i.e., the duration of inhibition) was under 15 seconds. Frequency of pulses was 60 Hz (rather than 10 Hz), and such stimulation never evoked sensations or linguistic utterances.

[3] A technology may eventually be developed to disrupt individual synapses, but currently the disruption of small groups of neurons is what is available (Tehovnik and Slocum 2007ab).

[4] Children learn verbal languages readily, but they require intensive study to read and write in a particular language. The symbols produced for writing (and for mathematics) are cultural inventions (Rojas et al. 2021).

[5] The human neocortex has a surface area of about 1,800 cm^2 (Van Essen et al. 2018). If the neocortex is composed of microcolumns measuring 30 μm in diameter (Peters and Sethares 1991), which would be the minimal size of a declarative, conscious unit devoid of its connections, then the neocortex should contain 64 x 10^6 declarative, conscious units, with each operating in parallel to encode a single feature (Logothetis et al. 2010; Murayama et al. 2011; 1991; Rutledge and Doty 1962; Schiller and Tehovnik 2015; Tehovnik and Slocum 2013). And each feature is stored according to context (Lu and Golomb 2023), which is determined by the connectivity profile per context. Each neuron in the neocortex has about 10,000 synaptic contacts, on average, suggesting an unlimited number of contexts per feature can be stored (estimate from chapter 18).