Contents

The thalamus is connected to the entire bottom layer of the cerebral cortex. It is the nexus of the various cortical processors as well as a recipient of independent input from most of the rest of the brain.

The thalamus is subdivided into numerous small and medium sized nuclei that between them receive inputs from every process in the nervous system (the white fibres in the illustration above largely penetrate the thalamus). The thalamic nuclei are interconnected which means that any of them could, potentially host activity from anywhere in the body or brain. Although the founders of neurology such as Hughlings Jackson and Penfield & Jasper located conscious experience in the diencephalon, including the thalamus, this is no longer the conventional wisdom. Processor Naive Realism is a prevalent philosophy in neuroscience and the small size of the thalamic nuclei means that they cannot support the processes that are assumed to compose consciousness. If the prejudice of Processor Naive Realism is abandoned the diencephalon and the thalamus in particular can be shown to be excellent candidates for the location of experience

(The Intralaminar Nuclei of the thalamus. The white space above and to the left of RN is the third ventricle. MD=mediodorsal nucleus. CM=Centromedian nucleus, RN=red nucleus (not part of thalamus) The black areas are stained white fibres). Picture linked from: http://www.neurophys.wisc.edu/neuro611/rayguillery/thalamus2002/Slide23.JPG )

 If the thalamus contains a location for conscious experience then lesions should abolish this experience. Unlike the cerebral hemispheres, lesions of the thalamus do indeed seem to abolish consciousness. The area that is most sensitive to lesions contains the Intralaminar Nuclei, especially the Parafascicular and Centromedian Nuclei. If these are damaged bilaterally patients suffer death, coma, akinetic mutism, hypersomnia, dementia and other equally serious impairments of consciousness that depend upon the size and placement of the lesions (Bogen 1995, Schiff & Plum 1999). In cases of fatal familial insomnia, in which patients exhibit many of these symptoms, there is marked neuron loss in the Intralaminar Nuclei (Budka 1998). The symptoms of bilateral damage to the ILN are often so severe that it is possible that the patients cease to be conscious and are being coordinated by automatic cortical processes.

Laureys et al (2002) investigated recovery from 'persistent vegetative state' (wakefulness without awareness). They found that overall cortical metabolism remained almost constant during recovery but that the metabolism in the prefrontal and association cortices became correlated with thalamic ILN and precuneus activity. Again confirming that thalamo-cortico-thalamic activity is required for consciousness and that cortical activity by itself is not conscious.

The Intralaminar Nuclei of the Thalamus are also called the "non-specific nuclei" because they do not serve a specific function and are not part of the Thalamic Relays that act as gateways for sensory data to the cerebral cortex.  As Bogen(1995) demonstrates, the ILN receive inputs, either directly or indirectly, from every part of the CNS but are "non-specific". What do they do?

Interest in the thalamus has recently been revived by the theories of Newman & Baars (1993), Baars, Newman, & Taylor1998) and Crick & Koch (1990). In Baars, Newman and Taylors' (1998) theory it is suggested that "The brain stem-thalamocortical axis supports the state, but not the detailed contents of consciousness, which are produced by cortex". They also propose that the "nucleus reticularis thalami" (Thalamic Reticular Nucleus, TRN), which is a thin sheet of neurons that covers the thalamus, is involved in a selective attention system. This concept is reinforced by the way that point stimulation of the TRN causes focal activity in the overlying cortex (MacDonald et al 1998) and the way the TRN is organised topographically (ie: has activity that is like an electrical image of receptor fields).

The thalamus is ideally placed for integrating brain activity, if tiny parts of the thalamus are removed consciousness is abolished and the thalamus is involved in attention and the global integration of cortical activity. Any impartial judge might pronounce that the site of conscious experience has been found, probably in the ILN of the thalamus, but no one can say how it works.

General anaesthesia should result in a profound depression of activity in the ILN if these are indeed the sites of the conscious state. White & Alkire (2003) administered halothane or isoflurane to volunteers and used positron emission tomography (PET) to monitor brain activity. They found severe depression of activity in the thalamus (Node 4 in the illustration below, red=most depressed). The depression appears from the diagram to be higher in the non-specific nuclei than in the relay nuclei of the thalamus. In other words the anaesthesia is neither turning off the cortex nor turning off the input to the cortex but it is turning off an important part of the thalamus. Fiset et al (1999) have also demonstrated a similar pattern of medial thalamic inactivity and cortical activity in propofol anaesthesia. Suppression of cortical activity is not the cause of unconsciousness; for instance, the anaesthetic agent chloralose leads to increased neural activity in the cortex relative to conscious patients (Cariani 2000).

Alkire et al (1996) also found that conscious recall memory of auditory stimuli was abolished by general anaesthesia whereas unconscious recognition memory continued. PET scans showed that activity in the Mediodorsal nucleus of the thalamus was required for conscious recall and was abolished by general anaesthesia. These experiments seem to confirm that cortical activity can occur without experience.

Sir Charles Wheatstone (1838) was the first scientist to systematically investigate binocular rivalry. Binocular rivalry occurs when different images are presented to the left and right eyes. The subject sees successively one image, a combined image and then the other image. The swapping of images can take a second or more. Binocular rivalry is of interest in consciousness research because the parts of the brain that contain the dominant image should also be those parts that are contributing to conscious experience. Binocular rivalry involves at least two components; the first switches from one image to a merged image and then to the other image and the second permits the view to be part of conscious experience.

The switching of one image for another may involve selecting one of the images as the percept or selecting one of the eyes. Blake et al (1979) performed an experiment in which subjects could change the image at a given eye by pressing a button. When a particular image became dominant they pressed a button to change the image at the eye receiving the dominant image for the non-dominant image. They found that the subjects immediately experienced the second image as the dominant image. This suggests that binocular rivalry is selecting between eyes rather than images. Lehky in 1988 proposed that the switching may be occurring as a result of feedback between visual cortical area V1 and the Lateral Geniculate Nucleus (a thalamic relay - see Carandini et al 2002) and Blake in 1989 also proposed that the switching occurred at the level of area V1. (Visual cortical area V1 receives visual input direct from the LGN.)

Tong (2001) has argued that, in humans, the switching of images in binocular rivalry may occur at the earliest levels in the visual cortex. In particular, Tong and Engel (2001) used an elegant technique measuring the activity in the visual cortex that represents the blind spot of the eye to show that almost complete switching to the dominant image occurs at the level of visual cortical area V1. In support of this idea of switching at the level of V1 or even before the cortex, Kreimann et al (2001, 2002) used direct electrode recordings in human cortex and found that the activity of most neurons changed with the percept.

In the case of binocular rivalry there appears to be a consensus that either cortical area V1 or the Lateral Geniculate nuclei are able to switch cortical input from one eye to the other. This results in all distal areas of cortex receiving input from only one source. This analysis is consistent with the experience of binocular rivalry in which the swapping of the images seems to occur automatically (non-consciously). It does not help us to pin down the location of conscious experience other than that it is distal to area V1 of the cortex (ie: entire cortex and areas receiving cortical output such as the thalamic Reticular nucleus and ILN etc.).

Pattern Rivalry is also of interest in consciousness research for the same reasons as binocular rivalry. In pattern rivalry a figure may have two or more forms that replace each other. Typical examples of such figures are the Necker cube and Rubin's face-vase. The similarity of the time course of the switching between percepts in binocular rivalry and pattern rivalry has led many authors to suggest that these involve the same mechanism. Logothetis et al (1996) used novel dichoptic stimuli (different images to each eyes) to produce a form of rivalry that seems to involve switching at levels in the cerebral cortex that are more distal to the sensory stimulus than V1. Leopold and Logothetis (1999), on the basis of their work with monkeys, state that "..many neurons throughout the visual system, both monocular and binocular, continue to respond to a stimulus even when it is perceptually suppressed.". Kleinschmidt et al (1998) investigated pattern rivalry with MRI and found activity in higher order visual areas during change of dominant pattern. Pettigrew (2001) also describes effects on rivalry due to thought and mood that may require involvement of large areas of cortex in the switching operation and stresses the way that V1 represents different visual fields in different hemispheres of the brain so that inter-hemispheric switching must also be considered.

It seems likely that the change of dominant pattern or percept is associated with higher level cortical activity but once the dominant percept is established many of the visually responsive neurons in the cortex are switched over to the new percept. This might account for the similarities in timing of binocular and pattern rivalry and the disparate results found by the various groups of authors. In the words of Kleinschmidt et al (1998):

"The transient activity fluctuations we found suggest that perceptual metastability elicited by ambiguous stimuli is associated with rapid redistributions of neural activity between separate specialized cortical and subcortical structures."

Which permits both the idea of selecting particular eyes or percepts, perhaps by feedback that switches a thalamic relay on the basis of cortical processing of patterns. Once the cortex has switched the thalamic relay most of the neurons in V1 would become exposed to the dominant percept but there would still be a few neurons in the cortical visual system receiving data from the non dominant image.

The investigations of binocular and pattern rivalry provide evidence that conscious visual experience is probably distal to V1 (ie: cortex or thalamus).

Perceptual rivalry may be part of complex decision making rather than being simply a switch to blank out unwelcome input. It is clear from the Rubin face-vase that pattern rivalry is linked to recognition and would involve a complex delineation of forms within cortical processing. This would suggest that many areas of cortex should be involved before a particular percept is made dominant. Pettigrew (2001) argues that rivalry is the result of a complex phenomenon rather than being simply a switching event. Pettigrew's discovery that laughter abolishes rivalry also points to a complex cortical system for switching percepts. Pettigrew proposes that complex cortical processes control rivalry and that the actual switching of percepts is performed sub-cortically in the Ventral Tegmental Area. He concludes his review of the problem by noting that "Rivalry may thus reflect fundamental aspects of perceptual decision making.." Pettigrew (2001).

Another effect, known as "binocular fusion", provides further compelling evidence for the non-conscious nature of the cerebral cortex. In binocular fusion images from both eyes are fused together to create a single image in experience. Moutoussis and Zeki (2002) used a form of binocular fusion in which images of faces were flashed at 100ms intervals to both eyes simultaneously. When both eyes received images of the same colour the subject could see the faces but when one eye received a green image on a red background and the other a red image on a green background the subjects reported seeing a uniform yellow field.

fMRI scans of the subject's brains showed that when both eyes were exposed to images of the same colour the part of the brain that deals with faces was active and when each eye received images of different colours the same areas of brain showed activity. In other words the cortex contained strong activity related to faces whether or not faces were experienced. Moutoussis and Zeki found a similar effect when they used images of houses instead of images of faces. The authors concluded that: "The present study further suggests that there are no separate processing and perceptual areas but rather that the same cortical regions are involved in both the processing and, when certain levels of activation are reached and probably in combination with the activation of other areas as well, the generation of a conscious visual percept".

This conclusion does not seem to be supported by the data. There is no evidence that any area of cortex contains the percept itself. The experiment shows that the cortex contains data relating to both red and green faces which suggests that the cortex is not the site of the conscious percept. The percept is most likely distal to the cortex in the thalamus.

It is interesting that Fries et al (1997) found that neurons that were activated by the dominant image in binocular rivalry fired synchronously whereas those that were activated by the non-dominant image did not. Thalamocorticothalamic oscillations are the most likely source for synchronising neurons over whole areas of cortex, again suggesting that the conscious percept is located in the thalamus rather than the cortex.

Our experience seems to contain entities with their attributes attached to them at the correct places in space and time. When a dog barks we see its jaws open at the same time as the bark and both jaws and bark are at the same location. We take this for granted but the brain must be engaging in some complex processing to achieve this synchronised and appropriately positioned set of objects and events. The illustration below shows the two basic processes that might be used to synchronise events between the different specialised processors in the cerebral cortex and brain in general.

In the first option a complete model of sensation, dream etc. may be created and then allowed to become part of conscious experience. In the second model events are released into experience as fast as possible but are synchronous when recalled, having been synchronised in a storage buffer. There is a third option in which there is no synchronisation of events so that the output from different processors would occur at different times.

The 'experience buffer' would be a volume of brain in which a succession of events could be recorded. The buffer might either be updated in steps, the previous content being discarded, or continuously updated with the oldest content being lost continuously.

In the first option events from different processes would always appear to be simultaneous unless the experience buffer were updated as a series of steps in which case any changes at around the moment of updating might appear in successive buffers. For instance, if change of position were processed before change in colour a circle on a screen that changed from green to red at the start of a motion might seem to be briefly green during the motion and then turn red.

In the second model events from different processors might appear asynchronous at the moment of experience but synchronous when recalled.

Colour vision and motion vision are processed in different parts of the visual cortex and in distinct parts of visual cortical areas V1 and V2. They are different processes and hence ideal for studying the synchronisation of cortical activity. Moutoussis and Zeki (1997) presented subjects with moving coloured squares on a computer screen that changed from red to green or vice versa as they changed direction of movement. It was found that subjects seemed to perceive changes in colour some 70-80 msecs before they perceived a change in the direction of motion of the squares. Further work by Arnold et al (2001) and Arnold and Clifford (2001) have confirmed that colour changes seem to be perceived before motion. Arnold and Clifford (2001) also found a quantitative relationship between the colour/motion asynchrony and the direction of change of motion, complete reversals of direction giving rise to the greatest asynchrony between the detection of colour and motion changes.

Moutoussis and Zeki (1997) conclude by stating that the asynchrony of neural processes shows that "..the perception of each attribute is solely the result of the activity in the specialised system involved in its processing..". It seems more likely that the experiments simply show that slow neural processes are not synchronised before they become percepts (the third option above). The experiments are excellent evidence for the concept of the cortex as a set of specialised processors that deliver their output asynchronously to some other place where the output becomes a percept.

These experiments on colour and motion suggest that there is no synchronisation between the processes that deal with these two aspects of vision. Another set of experiments by Clifford et al (2003) supports this idea of processing being asynchronous. They asked subjects to perform a variety of judgements of when visual events occurred and found that the degree of synchrony of one visual event with another depends on the type of judgement. Different judgements probably use processors in different areas of cortex and the output from these arrives asynchronously at the part of the brain that supports the percept.

When the percept is formed there must be feedback to the cortical processes that create its content. Otherwise it would not be possible to report about the percept and the cortex would be unable to direct processing to the percept in preference to other, non-conscious cortical data.

Although slow processes (20 milliseconds to 1 second) do not seem to be synchronised there is some evidence for very rapid synchronisation. Andrews et al (1996) revisited a problem raised by the famous physiologist Charles Sherrington. Sherrington considered the phenomenon of 'flicker fusion' in which a flickering light appears to be a continuous steady light if it flashes on and off at frequencies of about 45 Hz or higher. He reasoned that if the images from both eyes are brought together to form a single image then the frequency at which a flickering light appears to be steady should depend on whether one or two eyes are used. Flicker fusion should occur if each eye receives alternate flashes at only half the normal flicker fusion frequency. The flicker should disappear if the left eye receives flashes at 23 pulses per second and the right eye receives alternate flashes at 23 pulses per second. When Sherrington performed the experiment he found that this was not the case, using approximate figures, each eye required 46 pulses per second for fusion to occur. Sherrington proposed that the flicker fusion in alternate binocular presentation was occurring "psychically", outside of normal physiological processes.

Andrews et al duplicated Sherrington's result but investigated it further. They found that when lights were flashed in each eye alternately at low frequences (2 Hz) the experience was the same as a light being flashed in both eyes at this rate. At frequencies of four Hz and higher the subjects began to report that the lights being flashed alternately in both eyes seemed to flicker at the same rate as lights being flashed in both eyes at half the frequency. It seemed as if a flash in one eye followed by a flash in the other eye was being perceived as a single flash or "conflated" as the authors put it. The authors explained this effect by suggesting that the brain activity corresponding to the flashes was sampled for a short period and any number of flashes occurring during this period became perceived as a single flash. The maximum rate of sampling would be about 45 Hz. This idea is similar to option (1) above, where the buffer is filled and emptied 40 - 50 times a second.

An experience buffer that is refreshed at 40-50 times a second might also explain the results obtained with colour and motion asynchrony because synchronisation between processes may well happen too quickly to affect processes that occur at very slow rates. Singer and Gray (1995), Singer (2001) have proposed that synchronisation between neurones at about 45Hz is the discriminator between those neurones with activity that contributes to conscious experience and activity in other neurones. A rapid refresh rate in a sychronising buffer agrees with the results found by Fries et al (1997) in which visual cortical neurones that represent a percept underwent synchronous oscillations in the gamma frequency range (39-63 Hz). Tononi et al (1998) have also found synchronisation of neural activity in neurones that represent the percept.

 The gamma frequency oscillations in the cortex originate in the thalamus and are part of the 'arousal system'.

Under Construction

Bibliography and References