https://orcid.org/0000-0002-6928-4473
The purpose of this editorial is to highlight the innovative work of Arthur D. (Bud) Craig, Jr., who passed away this summer [Citation1]. His scientific contributions involve seminal studies of the neural substrates for temperature discrimination, thermal pain, thermal (dis)comfort, and thermoregulatory behavior that should be of great interest to the readers of this journal.
Bud was one of the few scientists whose work made a true difference. He was, with his own words, a functional neuroanatomist, working in the field of pain and temperature processing. Almost single-handedly he changed the concept of these modalities from being exteroceptive sensations to part of interoception, signaling information on the status of the body itself. He took this paradigm-changing idea to even further heights by identifying the previously largely neglected insular cortex as the primary sensory cortex for interoception, or the image of the “material me” that provides a homeostatic representation of the physiological condition of the body. He proposed and provided experimental evidence for the insula as the site critical for our awareness, and hence the potential substrate for our consciousness. I had the privilege to know him for over 40 years, and to collaborate with him since the late 1980s. He was the most brilliant person I have ever known.
Bud started his academic career with a major in mathematics from Michigan State University, but, as he describes in his monograph [Citation2], after having attended a class in functional neuroanatomy and watched the demonstration of the somatotopic representation of light touch in the sensory thalamus, he was hooked on neuroscience and began graduate studies in neurophysiology, neuroanatomy, and electrical engineering at Cornell University under the supervision of Daniel Tapper. After his PhD in 1978, with a thesis on sensory processing in the spinal cord, and in particular the spinocervical pathway, he moved to Washington University in St. Louis, MO, for postdoctoral work with Harold Burton. There he started what would be the theme for most of his career, studies on the organization of ascending pain and temperature pathways. As he also describes in his monograph, he could not understand why these pathways, while in textbooks being lumped together with the pathways for innocuous mechanoreceptive information from the skin and proprioception from muscles and tendons, took such a different route to the thalamus, crossing segmentally and running on the contralateral side of the spinal cord, and he decided to find out why. With Burton, Bud identified a specific termination site in the medial thalamus of the cat for spinothalamic fibers that originated in the superficial dorsal horn (lamina I), which was named nucleus submedius, and which did not receive input from deeper spinal cord laminae. It was suggested that this nucleus was a specific relay for nociceptive fibers (see for references to Bud Craig’s publications). (Later, with Jonathan Dostrovsky, Bud showed that the nucleus submedius also receives input from thermospecific lamina I neurons). This work was the first in a long series of studies from Bud showing that the spinothalamic neurons encoding pain and temperature and originating in lamina I (which make up about half the spinothalamic tract) constitute a distinct population with a distinct supraspinal termination pattern.
Table 1. Select publications by Bud Craig.
In 1981, Bud moved to Germany to join the Institute of Physiology directed by Robert Schmidt, first in Kiel and from 1983 in Würzburg. Working with Klaus-Dietrich Kniffki, Bud characterized the response properties of lamina I spinothalamic tract neurons. These neurons had only been sparsely recorded from because it was difficult to keep these superficially located cells in viable condition for the long-lasting experiments that were needed. Craig and Kniffki, using antidromic stimulation from the thalamus for identification of the lamina I projection neurons, found that these neurons could be separated into three distinct categories based on their response properties. One group of neurons responded selectively to peripheral nociceptive stimuli, in the form of pinch and/or noxious heat (nociceptive specific cells); a second group was selectively activated by cooling of the skin and inhibited by heat; and a third group responded not only to pinch and noxious heat but also to skin cooling (heat, pinch and cold; HPC cells). Later, in his own laboratory and together with postdoctoral fellow David Andrew, Bud would also identify lamina I spinothalamic tract cells responsive selectively to warm stimuli, and to the application of histamine, the latter finding hence demonstrating a central pathway specific for itch.
The HPC cells were initially called multireceptive cells, but Bud subsequently realized that they were indeed polymodal nociceptive, because their threshold for cooling stimuli was much higher than that of the thermoreceptive neurons. While the latter neurons responded to stimuli just below skin temperature, the HPC cells did not respond until the temperature was below that of the thermoneutral zone, i.e. about 24°C. The cool stimuli that excite them hence challenge homeostasis and are thus inherently noxious. For those who, like me, live in Northern Europe, it is well known that if you are out at sea, where even in the middle of summer the water temperature rarely reaches a little more than 20°C, and you fall overboard, you will only temporarily be saved by your floating device but eventually die of hypothermia.
The seminal study by Craig and Kniffki showed that the lamina I spinothalamic projection comprises physiologically distinct classes of cells that maintain the functional specificity of the small diameter afferent fibers, i.e. there are “labeled lines” to higher centers both for pain and temperature. Bud later also demonstrated that the three different classes of cells were morphologically distinct. These findings were at odds with the view held, and still held, by many investigators that pain is represented by patterned activity within a convergent somatosensory subsystem through so-called wide dynamic range neurons, mostly located in lamina V of the dorsal horn, which respond in a graded fashion to innocuous and then noxious stimuli. This idea is based on the so-called gate control theory that was introduced 1965, but shown soon after to be incorrect by, e.g., Robert Schmidt. Bud would later provide evidence that the nociceptive specific cells, dominated by Aδ input, and the polymodal HPC cells, dominated by C-fiber input, subserve the distinct sensations of first (sharp) and second (burning) pain, respectively. Furthermore, in a direct comparison of polymodal nociceptive lamina I spinothalamic tract neurons and lamina V wide dynamic range spinothalamic tract neurons, he showed by using a brief contact heat stimulus that in humans elicits a strongly augmented sensation of burning pain, that the response characteristics of the lamina I polymodal nociceptive STT neurons, but not the lamina V wide dynamic range STT neurons closely matched that of the human psychophysics. Unfortunately, this very elegant and important study has largely been neglected in the literature.
After Bud in 1986 had returned to the US to join the Barrow Neurological Research Institute in Phoenix, AZ (), he began to further characterize the projections from lamina I neurons, now including studies in monkeys, which the laboratory facilities at BNI enabled him to house and use. He showed that lamina I neurons project not only supraspinally, but also to the intermediate segment of the spinal cord, thus providing evidence for a direct pathway for the somatosympathetic reflex effects of innocuous thermal and noxious stimuli. Along the same vein, he showed that the spinal lamina I cells project to a variety of brainstem sites involved in autonomic (cardiovascular, respiratory) and homeostatic processing and the control of the behavioral state. He then set out to identify, in the primate, their thalamic target, a task which in his talks he jokingly compared to finding Waldo, the character in a children’s comic book (“Where is Waldo”) that was to be discerned in a very busy picture. While previous work had identified the target of the spinothalamic tract as being in the same relay nuclei as those receiving innocuous tactile information, Bud showed that the main termination field in primates is distinct from these nuclei and located in an area adjacent to the basal part of the ventromedial nucleus (VMb) that receives taste and visceral information from the nucleus of the solitary tract. He showed that this area, which he in analogy with the taste and visceral relay named the posterior portion of the ventromedial nucleus, or VMpo, contains neurons that similar to the spinothalamic lamina I neurons are selectively responsive to pain and temperate stimulation, and, very importantly, that it projects to the insular cortex, similar to the taste and visceral responsive neurons in the VMb. He showed that the insula contains a postero-anteriorly organized somatotopic map, and that it encodes both the somatic localization and the intensity discrimination of pain and temperature sensations. With this finding, all pieces fell into place: The lamina I spinothalamic tract neurons take their origin not from the alar plate, as do the somatosensory neurons, but from the intermediate gray matter, similar to the preganglionic autonomic neurons (this developmental history can be seen in birds where the superficial laminae are located lateral to the dorsal horn); they receive thin and not thick fibers, which enter via the Lissauer tract; they project via the contralateral dorsolateral funiculus and not ipsilaterally; and the body representation in the thalamus and insular cortex is organized orthogonally and not mediolaterally as in the somatosensory thalamus and cortex, i.e. the pathways for pain and temperature are massively distinct from those mediating innocuous tactile sensations. While the latter pathways are exteroceptive, providing mechanoreceptive information needed for musculoskeletal control, Bud realized that the pain and temperature pathways are interoceptive, signaling information about the status of the body itself, thus being closely related to the pathways that originate in the solitary tract nucleus and relay other sensations from the body such as taste, hunger, thirst, and visceral sensations. The concept of pain and temperature sensations as interoceptive signals that challenge homeostasis, and hence the integrity and survival of the organism, is consonant with the strong affective-motivational response they elicit. Those who have experienced severe pain know how it overshadows all other perceptions, and anyone who has come into a chilly house recalls how the cold penetrates to the core and the discomfort occupies the attention. Bud pointed out that while we normally regard temperature sensation as a discriminative cutaneous sensory capacity, the valence of the feeling depends on the body’s thermoregulatory needs [Citation6]. Hence, a glass of cool water is pleasant if you are overheated but aversive if you are chilled.
Figure 1. Bud Craig in his laboratory at Barrow Neurological Institute, Phoenix, AZ, in 2005. Courtesy of Lora Sanders.
Bud’s demonstration that the insular cortex was the primary sensory cortex for pain (and temperature) resolved the long-standing paradox that, as was pointed out already by Head and Holmes in 1911, lesions of the convexity of the cortex, including the somatosensory cortex, while producing extensive sensory loss failed to abolish the sensation of pain. It also explained why pain rarely was elicited by the stimulation of the cortical surface in the classical studies done by Penfield in the 1930s; when the Penfield experiment was repeated some 10 years ago but the stimulated area now included the insular cortex, pain was elicited and was somatotopically represented [Citation3], as Bud had postulated.
The demonstration of a primate-specific spino-thalamo-insular pathway for pain and temperature as well as for other sensations from the body itself, represented a paradigm shift, and was a scientific contribution well on par with the discovery of the molecular mechanisms of these sensations that recently was rewarded with the Nobel prize (https://www.nobelprize.org/prizes/medicine/2021/press-release/). However, since it challenged the prevailing ideas, it evoked fierce opposition among many of Bud’s contemporaries who had devoted a large part of their careers to those very ideas. While the criticism from several was done in good faith and the following discussion helped clarify important points, that from others was colored by personal animosity with attempts to discredit Bud and his work. Respected journals published pieces with derogatory titles such as “Pain in the brain and lower parts of the anatomy,” and “A pain in the thalamus.”
A recurrent argument against Bud’s findings was that the VMpo did not exist and that the activation of the insular cortex by nociceptive stimuli was relayed via the S1 and S2 cortices. Unfortunately, Bud did not get to publish his extensive data on the VMpo to insula projection until many years later, and although he had reported on them in several book chapters, the findings were deliberately neglected. Bud’s delay in publishing his data was in part due to his shift of focus from pain and temperature to a global understanding of the function of the interoceptive information in the insula, and in particular its role for human awareness. However, I also got the impression he was tired and disappointed that his work in the pain field did not get the recognition it deserved. Sadly, even today Bud’s breakthrough studies are often ignored: To read the account of the central pain pathway in one of the more authoritative neuroscience textbooks [Citation5] is a surreal experience; and although the pain pathway is correctly shown to terminate in the insula in the chapter on the sensation on touch, Bud’s work is not mentioned.
The presence of pain- and temperature-specific pathways from the dorsal horn to the insula, intermixed with neurons selectively responsive to stimuli that produce various bodily feelings in humans, including itch, sensual touch, and muscle and visceral sensations, as well as the representation in the insula of vagal-responsive and gustatory neurons, made Bud to postulate that the insula harbored the primary sensory cortex for interoception, or the image of the “material me” that provides a homeostatic representation of the physiological condition of the body. He postulated that the interoceptive feelings are re-represented, and multimodally integrated, in more anterior portions of the insula in sequence of increasingly homeostatically efficient representations that integrate all salient neural activity. In a seminal study with imaging technique, Bud showed that human brain activity correlated with graded cooling stimuli in the contralateral dorsal margin of the middle/posterior insula (i.e. in the primary interoceptive cortex), whereas perceived thermal intensity was correlated with activation in the right (ipsilateral) anterior insular cortex. He suggested that subjective awareness is associated with activation of the anterior insular cortex and supported these ideas in studies of patients with epileptic seizures with ecstatic auras. He demonstrated in human volunteers that the insula is critical for fluid intelligence and the perception of time. He conceptualized these findings in several influential, very highly cited review articles [Citation4,Citation7], in which he provided a unified theory, with feelings (and awareness) engendered in the insular cortex, and motivations (and agency) engendered in the cingulate cortex. He integrated these ideas into the role of the bicameral forebrain, in that sympathetic activity, negative affect, avoidance behavior and energy expenditure are subserved predominantly by the right forebrain while parasympathetic activity, positive affect, approach behavior and energy nourishment are subserved predominantly by the left forebrain, with opponent interactions between the two sides.
While, as mentioned, Bud’s work on the central pathways for pain and temperature was met by many with disbelief, his studies on the role of the insular cortex for interoception were very well received, and propelling a very extensive research into its involvement in, e.g., drug addiction, alcoholism, anxiety, and eating disorders, work to which Bud also contributed. This positive response may seem a bit paradoxical since the concept of interoception in the insula was built on his work on the lamina I spinothalamocortical projection that others challenged, but is probably explained by that his ideas on interoception and awareness had a strong explanatory value and created the framework which many in that field, consisting of researchers other than those in the field of pain, had been looking for.
Many scientific discoveries are based on novel tools and techniques, as implied in the pun “the gain in the brain is in the stain.” However, major paradigm-changing ideas are the result of the cognitive processes of brilliant minds. Bud’s contributions clearly belong to the latter category. While he strived to have the best possible equipment, he worked with established neuroanatomical and neurophysiological methods, although he used them in an unprecedented way and with utmost rigor. Later, he also employed functional imaging techniques. While he certainly had an advantage over many in that he mastered both electrophysiology and neuroanatomy, his accomplishments were the results of his ability to ask the right questions, to see the larger picture in the details, and to interpret the findings against an encyclopedic knowledge of the literature. Much of what he wrote during his last active years was of a theoretical nature, in which his vast knowledge was used to formulate fascinating hypotheses. One such was his idea about how time is perceived and represented; its relation to music; and why time feels to expand and contract depending on the emotional state. Reading his article on the topic shows his extraordinary grasp of the literature, and his ability to formulate thrilling and yet testable hypotheses.
One of Bud’s strengths was that he always went back to the original literature, even if it was papers published in the nineteenth century. He was not satisfied to read what others reported on previous work, but to get to know himself what actually was written. He once asked Ulf Norrsell, a physiologist in Gothenburg with whom he studied thermosensory pathways in the cat, if there was any known pain illusion. Ulf directed Bud to an experimental psychology textbook from the 1940s, in which there was a reference to such an illusion in a study published in 1896 by a Swedish physiologist, Torsten Thunberg. In this illusion a painful heat sensation is elicited by touching interlaced warm and cool bars to the skin. Thunberg’s study was published in Swedish in a local university journal and while Bud was well versed in German, he did not master Swedish, so I got a phone call from him asking me to provide a translation, which I did (this was before machine translations became available). It was not long until he had constructed what he called a thermal grill, and explained the mechanisms behind the illusion, with the suggestion that the same mechanisms lay behind the cold-evoked burning pain seen in central pain syndromes.
The thermal grill illusion shows how pain and temperature sensations interact, or more precisely how cold inhibits pain, a phenomenon Bud was interested in. It is an everyday experience that cooling a damaged body part alleviates the pain, and although this effect may involve desensitization of peripheral nociceptors it most likely also engages central mechanisms, because the application of cold has been reported to reduce also pain that is evoked by electrical stimulation of a peripheral nerve. Bud’s interest in the thermal grill illusion followed on his finding that nociceptive and thermoreceptive lamina I spinothalamic neurons displayed opposing responses to morphine, suggesting antagonistic functions: Whereas most nociceptive neurons were inhibited by systemic morphine, the responses of the thermoreceptive selective neurons were enhanced. In the thermal grill illusion, touching interlacing warm (40°C) and cold (20°C) bars elicits the sensation of burning pain, whereas touching each bar separately only evokes the sensation of warm and cold, respectively. Electrophysiological recordings from lamina I spinothalamic neurons showed that whereas the cold stimulus excited both polymodal nociceptive and thermoreceptive specific neurons, the response of the thermoreceptive specific neurons, but not that of the polymodal nociceptive neurons, was attenuated by the addition of the warm stimulus, probably via spatial summation. Bud explained the thermal grill illusion as disinhibition of the polymodal nociceptive neurons: Under normal conditions the thermoreceptive specific neurons inhibit the activity of the polymodal nociceptive neurons, and when this inhibition is mitigated, pain appears. Bud further suggested that such impaired inhibition could be the basis for central pain, elicited by imbalanced forebrain integration of pain and temperature activity. Indeed, a pathognomonic feature of central pain is loss of thermosensibility.
Bud’s scientific accomplishments are extraordinary, and they become even more extraordinary considering that he never had a large laboratory but worked on his own or with a small number of lab members like a single postdoc and a technician. Of the close to one hundred papers he published after he had established his own laboratory in Phoenix, an astonishing quarter is with him as single author, and includes many extensive and important studies with primary data. His papers, including those done at the beginning of his career are still a pleasure to read. The writing is precise, they all contain gems of insight, and especially those that report his work on monkeys represent a treasure, because it is very unlikely that something comparable ever will be done again.
Bud was a very intense person and as such was a charismatic lecturer (https://vimeo.com/8170544?login=true). Working with him was exciting but also very demanding because he strove for perfection. When we were finalizing our manuscript on the VMpo nucleus in humans, we spent weeks with the microscope, discussing the fine details of cytoarchitectonic borders. When something caught Bud’s interest, be it in science or other fields, he easily became engulfed. When I once took him out on rod fishing on the Baltic coast, he was first bored because nothing happened but after the fish had started biting and the first pike had been landed in the boat, he never wanted to stop. After that, joint fishing expeditions were on the schedule during my visits to Phoenix. While he was a very warm person and very loyal to his friends, he had an uncompromising attitude to scientific work. He despised poor science, and had little understanding for the politics around papers, where a study that was neither very well done nor conclusive never-the-less was published because someone, be it a postdoc or graduate student, “needed it.” He was frank, and often not very diplomatic in his criticism, which unfortunately gave him several adversaries. He was, rightly, disappointed that his work did not get the recognition it deserved. But he was also open to criticism. If someone had a valid point, Bud went back to the laboratory and carried out the necessary experiments. He was, in every sense of the word, a true scientist.
In his last years, Bud was sadly afflicted by illness. He was struck with multiple sclerosis and came to suffer from central pain, a condition his research had helped explain. While his splendid intellect remained intact, his powers waned significantly, and he had to retire early. Lots of important data he had collected remain unpublished. Hopefully, his recent collaborators will be able to get them out. I also hope that he gets the recognition he so clearly deserves, even though it be posthumously.
The author thanks Drs. Jonas Broman, Jonathan Dostrovsky, David Engblom, Henry Evrard, and Andrej Romanovsky for comments and suggestions on earlier versions of the text.
References
- Blomqvist A, Evrard HC, Dostrovsky JO et al. A. D. (Bud) Craig, Jr. (1951–2023). Nat Neurosci. 2023. doi: 10.1038/s41593-023-01463-9
- Craig AD. How do you feel?: An interoceptive moment with your neurobiological self. Princeton, NJ: Princeton University Press; 2014. p. 384.
- Mazzola L, Isnard J, Peyron R, et al. Stimulation of the human cortex and the experience of pain: Wilder Penfield’s observations revisited. Brain. 2011;135(2):631–640. doi: 10.1093/brain/awr265
- Craig AD. How do you feel? Interception: The sense of the physiological condition of the body. Nat Rev Neurosci. 2002;3(8):655–666. doi: 10.1038/nrn894
- Kandel ER, Koester JD, Mack SH, et al, editors. Principles of neural science. 6th. New York, NY: McGraw Hill, 2021, p. 1646
- Craig AD. Central neural substrates involved in temperature discrimination, thermal pain, thermal comfort, and thermoregulatory behavior. Handb Clin Neurol. 2018;156:317–338.
- Craig AD. How do you feel — now? The anterior insula and human awareness. Nat Rev Neurosci. 2009;10(1):59–70. doi: 10.1038/nrn2555