Turning to the other property of grandmother cells, convergence, the most extreme example of neural convergence is William James’s concept of a “pontifical cell” whose activity is identical to consciousness, as in this passage from his Principles of Psychology:
There is, however among the cells one central or pontifical one to which our consciousness is attached. But the events of all the other cells physically influence this arch-cell; and through producing their joint effects may be said to “combine.”27
C.S. Sherrington in his classic Man on his Nature took up James’s ideas that there might be “convergence . . . of the nervous system . . . onto one ultimate “pontifical nerve-cell.”28 He then rejected pontifical cells in favor of an ensemble cell theory of consciousness as “a million-fold democracy whose each unit is a cell.”
Barlow thought that the proposal of grandmother cells was inadequate because the multidimensional aspects of visual percepts could not be represented by a single individual cell.29 Rather, he proposed that a small number of cells would be needed to represent a percept. He named such cells “cardinal” cells since cardinals are lower in the hierarchy than popes and there are more of them.
CONCLUDING COMMENT
The idea that there might be convergence of neural input onto a single cell which would provide that cell with the ability to represent a complex and specific percept seems to have arisen independently several times, first as the gnostic cells elaborated in detail by Konorski and then as the grandmother cells deriving from Lettvin’s parable. Cells with properties that are similar to those of gnostic and grandmother cells have been found in both the inferior temporal cortex and the hippocampus. Grandmothers (and other complex objects) may be represented by ensembles of “grandmother” cells that vary in their responses to different aspects of the stimulus. Finally, contemporary human brain imaging studies have yielded specialized regions of the cortex that closely resemble the gnostic fields proposed by Konorski.
POSTSCRIPT
Since our discovery of neurons in monkeys that respond much more strongly to images of faces than to other images or to scrambled faces30 there has been an enormous increase in our understanding of these neurons. We now know that some of these neurons are more sensitive to facial identity, some to emotional expression of faces,31 and some to the gaze position of the eyes.32
FACE-SELECTIVE NEURONS
Face-selective neurons are found throughout inferior temporal cortex but are concentrated in two regions of the superior temporal sulcus, located on the dorsal border of the inferior temporal (IT) cortex.33 They are also found immediately dorsal to IT (in the superior temporal polysensory area), in the amygdala, and in the ventral prefrontal cortex.34 The face-selective cells in these different areas are thought to play different roles in the processing of the facial image.
Face-selective cells may be selective for face orientation or may generalize across different lateral orientations.35 The response to inverted faces is much reduced.36 As is usually the case with other IT neurons, their response selectivity remains invariant over changes in size, contrast, and exact retinal location.37
Furthermore, face-selective cells, at least in IT cortex, do more than process the retinal image; they have other cognitive functions as well. Like other IT neurons, they can be strongly modulated by attention and have mnemonic properties in both short and long time scales.38 Although probably present at birth, they can be modified with experience.39
Overall these properties of face-selective cells suggest strongly that they may be involved in the behavioral detection and discrimination of faces. This idea is confirmed by the demonstration that electrical stimulation of a cluster of IT cells elicits face recognition.40
Two early generalizations made in the chapter remain valid. The first is that such apparent gnostic neurons have been found only for faces and body parts and not for other natural stimuli important for a monkey. The second is that face-selective cells do not fit strict criteria of “grandmother cells.” Rather than single, dedicated grandmother cells, an ensemble of face-selective cells is required to represent a specific face.
FACE PROCESSING IN HUMANS
Whereas it took about 12 years for anyone to try and replicate our basic findings on IT neurons in monkeys, it took about 19 years for the search for responses to faces in the temporal cortex of humans to begin. Selective responses to faces were finally shown in human ventral temporal cortex, especially the fusiform gyrus, by PET scanning, single-unit recording, field potentials, and fMRI.41 Since then there has been a flood of imaging papers on face processing in the human temporal lobe.
Functional magnetic resonance imaging (fMRI) has revealed three main face-selective patches in humans: the occipital face area (OFA), the superior temporal face area (STS-FA), and the most studied and first discovered fusiform face area (FFA). The OFA has been thought to be involved in processing face components; the FFA in processing face identity; and the STS-FA in processing gaze direction.42 Recently up to two anterior face-selective regions have been described, perhaps involved in face recognition.
There are two current questions about the FFA. The first is whether the FFA is a “pure” face area or an “expertise” area; the current evidence suggests that within the FFA there are exclusive face-processing mechanisms. The second question is on the role of the widespread distributed processing of the facial image throughout human ventral cortex outside of the three “focal” areas: what are the relative roles of the focal and distributed processing of faces?
MONKEY NEURONS AND HUMAN FACE-SELECTIVE PATCHES
In a beginning attempt to bridge monkey single-unit and human fMRI studies, we and others have used fMRI to demonstrate several face-selective patches in monkey temporal cortex.43 Tsao and her colleagues showed that in a face-sensitive patch (as detected using fMRI) in the monkey that may correspond to human FFA, virtually all the single neurons recorded were face selective.44 Subsequent fMRI studies showed that there were at least five selective face patches in the monkey temporal lobe,45 the largest two corresponding to the areas in which previous researchers had found the highest concentration of face-selective cells. Electrical stimulation of four of the face patches revealed by fMRI showed that they were interconnected, forming a “specifically interconnected hierarchical network.”46 Furthermore, both Tsao et al. and Pinsk et al. showed that the multiple face patches in monkeys corresponded closely to the multiple face patches in humans.47 Beyond beginning to establish homologies between humans and monkeys, further elucidation of the temporal face-processing network is likely to throw light on the circuitry underlying not only face processing but also other types of pattern recognition.
A CANDIDATE FOR TRUE GRANDMOTHER CELLS?
As noted above, it was generally agreed that inferior temporal cells coded faces by small ensembles of cells and rather than by dedicated “grandmother cells.”48 Similarly, there were an increasing number of examples from other systems where a small number of cells (or “sparse coding”) were sufficient to code complex phenomena, such as place cells in rat hippocampus49 or time in HVC neurons of the song system in songbirds.50
Then in 2005 a group at Cal Tech reported cells in the human medial temporal lobe (hippocampus and several adjacent areas) that responded with astonishing invariance and specificity.51 For example, one responded only or best to various pictures of the actress Halle Berry including images of her dressed as a “cat woman” in a screen role of hers. “Notably,” to quote the authors, “it was selectively activated by the letter string ‘Halle Berry’.” The cell did not respond to other faces, cat women, or letter strings (see figure 12.4). Another cell responded best to various pictures of the Sydney opera house and the letter string “opera house” but not to a variety of other buildings and letter strings. Both Ms. Berry and the opera house were familiar to the subject, and these stimuli were chosen for further analysis because in preliminary testing they had elicited responses. These results were certainly much closer to il
lustrating the idea of a grandmother cell than any previous ones.
Figure 12.4
A single unit in the right anterior hippocampus activated by different views of the actress Halle Berry and not by a variety of other faces. The cell also responds to a drawing of her and to herself dressed as Catwoman in a recent movie and to the letter string of her name. The vertical dashed lines mark stimulus onset and offset, which were 1 second apart (Quiroga et al., 2005). Reprinted by permission from Macmillan Publishers Ltd.
In a subsequent paper the authors make it clear that, for several reasons, they had found “sparse but not ‘Grandmother-cell’ coding.”52 First, it was simply improbable that a single cell would show invariant responses to only one individual, especially as the stimulus set tested was small. Second, in fact some cells responded to two individuals (on the same TV program) or to two sites (seen on a recent trip). Third, the usual response latency (250–350 ms) was much longer than typical recognition times. The second and third points suggested that these cells might play a role in memory for abstract concepts rather than in sensory processing.
NOTES
This chapter is adapted from an article published in The Neuroscientist (8: 512-518 [2002], “The genealogy of the ‘grandmother cell’”).
1. Barlow, 1972; Blakemore, 1973b; Anstis, 1975; Frisby, 1980; Marr, 1982; Churchland, 1986.
2. Rosenzweig et al., 1999; Gazzaniga et al., 1998; Cowey, 1994.
3. Lettvin, personal communication in Barlow, 1995.
4. Roth, 1969
5. Barlow, 1972
6. Blakemore, 1973b.
7. Konorski, 1967
8. E.g., Martin et al., 2000; Caramazza, 2000.
9. Hubel and Wiesel, 1962, 1965.
10. Hubel and Wiesel, 1965.
11. Mishkin, 1966.
12. Allman and Kaas, 1971.
13. Pribram and Mishkin, 1955.
14. Konorski, 1967.
15. Gross, 1968.
16. Gross, Bender, and Rocha-Miranda, 1969; Gross, Rocha-Miranda, and Bender, 1972; Gross, 1994.
17. Lettvin et al, 1959, 1961.
18. Perrett, Rolls, and Caan, 1982; Rolls, 1984; Yamane, Kaji, and Kawano, 1988.
19. Fonberg, 1974; Konorski, 1974.
20. Logothetis and Sheinberg, 1996; Tanaka, 1996.
21. Gross, 2000b; Kreiman, Koch, and Fried, 2000, 2001.
22. Desimone, 1991; Gross, 1992.
23. Gross, 1998a.
24. Boring, 1950.
25. Müller, 1965.
26. Adrian and Matthews, 1927.
27. James, 1890.
28. Sherrington, 1940.
29. Barlow, 1972.
30. Gross, Rocha-Miranda, and Bender, 1972.
31. Hasselmo, Rolls, and Baylis, 1989.
32. Perrett, Rolls, and Caan, 1982
33. Tsao and Livingstone, 2008, figure 8.
34. Bruce and Gross, 1981 (dorsal to IT); Leonard, Rolls, and Baylis, 1985 (amygdala); Scalaidhe et al., 1999 (ventral prefrontal cortex).
35. Desimone, Albright, and Gross, 1984.
36. Perrett et al., 1985.
37. Desimone, Albright, and Gross, 1984; Schwartz et al., 1983.
38. E.g., Miyashita, 1988, Miller, Li, and Desimone, 1991; Colombo and Gross, 1994; and Desimone, 1996.
39. Rodman, O’Scalaidhe, and Gross, 1993; Logothetis, Pauls, and Poggio, 1995.
40. Afraz, Kiani, and Esteky, 2006.
41. Haxby et al., 1991; Sergent and Signoret, 1992; Ojemann, Ojemann, and Lettich, 1992; Allison et al., 1994; Kanwisher, McDermott, and Chun, 1997; Puce, Gore, and McCarthy, 1995.
42. See references in the review by Tsao and Livingstone, 2008.
43. Tsao et al., 2003; Pinsk et al., 2005; Logothetis et al., 1999.
44. Tsao, Freiwald, and Tootell, 2006.
45. Tsao, Moeller and Freiwald, 2008; Pinsk et al., 2009
46. Moeller, Freiwald, and Tsao, 2008.
47. Tsao, Moeller and Freiwald, 2008; Pinsk et al., 2009
48. Desimone, 1991; Gross, 1992; Abbott, Rolls, and Tovee, 1996; Rolls and Tovee, 1995.
49. Jung and McNaughton, 1993; Thompson and Best, 1989.
50. Hahnloser, Koshevnikov, and Fee, 2002.
51. Quiroga et al., 2005.
52. Quiroga et al., 2007.
REFERENCES
Abbott, L. F., Rolls, E. T., and Tovee, M. J., 1996. Representational capacity of face coding in monkeys. Cereb. Cortex, 6: 498–505.
Abrous, D. N., Koehl, M., and Le Moal, M., 2005. Adult neurogenesis: From precursors to network and physiology. Physiol. Rev., 85: 523–569.
Adair, H., and Bartley, S. H., 1958. Nearness as a function of lateral orientation in pictures. Percept. Mot. Skills, 8: 135–141.
Adams, K. R., 1969. Reflections and laterality. Leonardo, 2: 301–307.
Adrian, E. D., and Matthews, B., 1927. The action of light on the eye. Part 1. The discharge of impulses in the optic nerve and its relation to the electric changes in the retina. J. Physiol., 97: 378–414.
Afraz, S. R., Kiani, R., and Esteky, H., 2006. Microstimulation of inferotemporal cortex influences face categorization. Nature, 442: 692–695.
Alhazen, 1989 [ca. 1021]. The Optics of Ibn al-Haytham. Books I–III: On Direct Vision. A. I. Sabra, trans. Warburg Institute, London.
Alhazen, 2001 [ca. 1021]. Alhacen’s Theory of Visual Perception: A Critical Edition, with English Translation and Commentary, of the First Three Books of Alhacen’s “De Aspectibus,” the Medieval Latin Version of Ibn al-Haytham’s “Kitab al-Manazir.” A. M. Smith, trans., ed. American Philosophical Society, Philadelphia.
Alhazen, 2006 [ca. 1021]. A. M. Smith, trans., ed. Alhacen on the Principles of Reflection: A Critical Edition, with English Translation and Commentary, of Books 4 and 5 of Alhacen’s “De Aspectibus,” the Medieval Latin Version of Ibn al-Haytham’s “Kitab al-Manazir.” American Philosophical Society, Philadelphia.
Allen, E., 1912. The cessation of mitosis in the central nervous system of the albino rat. J. Comp. Neurol., 19: 547–568.
Allison, T., Ginter, H., McCarthy G., Nobre, A. C., Puce, A., Luby, M., and Spenser, D. D., 1994. Face recognition in human extrastriate cortex. J. Neurophysiol., 71: 821–825.
Allman, J. M., and Kaas, J. H., 1971. A representation of the visual field in the caudal third of the middle temporal gyrus of the owl monkey (Aotus trivirgatus). Brain Res., 31: 85–101.
Alting, M.P.C. and Waterbolk, T. W., 1982. New light on the anatomical errors in Rembrandt’s Anatomy Lesson of Dr. Nicolaas Tulp. J. Hand Surg. 7: 632–634.
Altman, J., 1962. Are new neurons formed in the brains of adult mammals? Science, 135: 1127–1128.
Altman, J., 1963. Autoradiographic investigation of cell proliferation in the brains of rats and cats. Anat. Rec., 145: 573–591.
Altman, J., 1967. Postnatal growth and differentiation of the mammalian brain, with implications for a morphological theory of memory In: Quarton, G. C., Melnechuck, T., and Schmitt, F. O., eds., The Neurosciences: A Study Program. Rockefeller University Press, New York.
Altman, J., 1969. Autoradiographic and histological studies of postnatal neurogenesis. IV. Cell proliferation and migration in the anterior forebrain, with special reference to persisting neurogenesis in the olfactory bulb. J. Comp. Neurol., 137: 433–458.
Altman, J., 2009. The discovery of adult mammalian neurogenesis. In: Seki, T., ed. Neurogenesis in the Adult Brain. Springer, Berlin.
Altman, J., and Bayer, S. A., 1995. Atlas of Prenatal Rat Brain Development. CRC Press, Boca Raton, FL.
Altman, J., and Bayer, S. A., 1996. Development of the Cerebellar System: In Relation to Its Evolution, Structure, and Functions. CRC Press, Boca Raton, FL.
Altman, J., and Das, G. D., 1965. Autoradiographic and histological evidence of postnatal hippocampal neurogenesis in rats. J. Comp. Neurol., 124: 319–335.
Altman, J., and Das, G. D., 1966a. Autoradiographic and histological studies of postnatal neurogenesis. I. A longitudinal investi
gation of the kinetics, migration and transformation of cells incorporating tritiated thymidine in neonate rats, with special reference to postnatal neurogenesis in some brain regions. J. Comp. Neurol., 126: 337–390.
Altman, J., and Das, G. D., 1966b. Autoradiographic and histological studies of postnatal neurogenesis. II. A longitudinal investigation of the kinetics, migration and transformation of cells incorporating tritiated thymidine in infant rats, with special reference to postnatal neurogenesis in some brain regions. J. Comp. Neurol., 128: 431–474.
Anderson, P., Morris, R., Amaral, D., Bliss, T., and O’Keefe, T., eds., 2006. The Hippo-campus Book. Oxford University Press, Oxford.
Anonymous, 1881a. The festivities of the congress. Br. Med. J., 2: 303–304, 752, 822– 824, 836–842.
Anonymous, 1881b. The charge against Professor Ferrier under the Vivisection Act: Dismissal of the summons. Br. Med. J., 2: 836–842.
Anstis, S. M., 1975. What does visual perception tell us about visual coding? In: Gazzaniga, M. S., and Blakemore, C., eds., Handbook of Psychobiology. Academic Press, New York.
Aretaeus, 1856 [2nd C.]. The book of cures. In: Adams, F., ed. and trans., The Extant Works of Aretaeus, the Cappadocian. Sydenham Society, London.
Arnheim, R., 1974. Art and Visual Perception. University of California Press, Berkeley.
Arnott, R., Finger, S., and Smith, C. U. M., eds., 2003. Trepanation: History, Discovery, Theory. Swets and Zeitlinger, Lisse, Netherlands.
Asanuma, H., 1975. Recent developments in the study of the columnar arrangement of neurons within the motor cortex. Physiol. Rev., 55: 143–156.
A Hole in the Head Page 23