Miller (1994) points out that ignition of a widely distributed but nevertheless tightly connected population of cortical neurons (cell assembly) may, in principle, lead to spectral responses in various frequency bands. However, if recent quantitative neuroanatomical data about fiber size frequencies in humans are valid, cortico-cortical information exchange takes place primarily in the millisecond range. If reverberations in cortico-cortical loops underlie differential spectral responses, they may accordingly be expected to be visible primarily in the gamma range. Miller is absolutely right in pointing out that the model presented in our target article is only one way of accounting for the data. At this point, the development of competing models appears to be most desirable in the field of cognitive neuroscience.
2. In their article on the fiber composition of the human corpus callosum, Aboitiz, Scheibel, Fisher and Zaidel (1992) write that "unmyelinated fibers were scarce, except in the genu (of the callosum), where they were found to comprise about 16% of the total fibers... In the rest of the callosal regions, there were almost no unmyelinated fibers (usually less than 5%)" (p.148). This contrasts to findings in other animals, where slow unmyelinated axons appear to be as numerous as fast myelinated ones (references in Miller, 1994). Miller points out that the data reported by Scheibel's group may be biased in favor of myelinated fibers. However, in the rhesus monkey, the percentage of unmyelinated fibers has been found to be almost as low as in humans (7-20%) (LaMantia & Rakic, 1990). This suggests that myelinated fibers are much more relevant for cortico-cortical information exchange compared to unmyelinated ones in monkeys or humans (while the opposite may apply for other animals). The quantitative analysis by Aboitiz et al. indicates that myelinated fibers of calibers 0.5-1.5 um are by far the most frequent axon type (similar results for monkeys; LaMantia & Rakic, 1990). If such fibers form a loop from Broca's to Wernicke's region and back, the round-trip time in such a loop will be some 20-40 ms, equaling a circulation frequency of 25-50 Hz. Thus, the basic argument is the following: there are quantitative data suggesting that fast axons are much more powerful than slow ones in the human and monkey cortex. Therefore, circulation of activity should lead to power changes primarily in upper (gamma) frequency bands. Of course, Miller is right in pointing out that there is a wide range of axon calibers and this makes fast as well as slow circulation possible.
3. In our target article, we proposed that each word is represented in the cortex as a transcortical cell assembly. According to our model, individual neurons or local neuronal clusters of such an assembly may correspond to phonetic, syntactic and semantic features of the word (Braitenberg & Pulvermueller, 1992; Pulvermueller, 1992). Assemblies may overlap if the words they represent share phonetic, syntactic or semantic features. So the assumption is that words are represented in distinct but possibly overlapping assemblies. The data reported so far do not prove that neuronal equivalents of different words are distinct. This is still a theoretical assumption. However, it has already been shown that words of different lexical categories elicit evoked potentials with distinct topographies (Neville et al., 1992; Pulvermueller et al., 1994b). Future research will possibly provide further evidence that different words have different cortical representations.
4. Miller asks for neuroanatomical evidence for projections from the anterior to the posterior perisylvian cortex and back. Pandya and Yeterian (1985) and, more recently, Deacon (1988; 1992) summarize data from the macaque monkey indicating that the possible homologues of Broca's and Wernicke's regions are strongly connected to each other (while the primary motor and auditory cortices appear to lack strong projections to each other). Given that there are transcortical assemblies including neurons of the perisylvian region, these data suggest that they are held together through fibers projecting from Broca to Wernicke and vice versa.
5. I fully agree with Miller's call for more detailed network explanations of differential gamma-band responses. The simple networks presented in our target article can only serve as models from which very general properties of cortical dynamics may be inferred. The processes of reverberation in cortico-cortical loops and of oscillation due to local inhibitory circuits are only two candidate processes leading to an account for different gamma-band responses evoked by words and pseudowords. Additional models that may also account for the data are most welcome.
Aboitiz, F., Scheibel, A.B., Fisher, R.S. & Zaidel, E. (1992) Fiber composition of the human corpus callosum. Brain Research 598:143-153.
Braitenberg, V. & Pulvermueller, F. (1992) Entwurf einerneurologischen Theorie der Sprache. Naturwissenschaften 79:103-117.
Deacon, T.W. (1988). Human brain evolution: 1. evolution of language circuits. In Intelligence and evolutionary biology. H.J. Jerison & I. Jerison, (Eds.). Berlin: Springer Verlag.
Deacon, T.W. (1992) Cortical connections of the inferior arcuate sulcus cortex in the macaque brain. Brain Research 573:8-26.
LaMantia, A.S. & Rakic, P. (1990) Cytological and quantitative characteristics of four cerebral commissures in the rhesus monkey. Journal of Comparative Neurology 291:520-537.
Miller, R. (1994) Cognitive Processing, But Not Cell Assembly Ignition. Commentary on Pulvermueller et al. on Brain-Rhythms. PSYCOLOQUY 5(50) brain-rhythms.2.miller.
Neville, H.J., Mills, D.L. & Lawson, D.S. (1992) Fractionating language: different neural subsystems with different sensitive periods. Cerebral Cortex 2:244-258.
Pandya, D.N. and Yeterian, E.H. (1985). Architecture and connections of cortical association areas. In Cerebral cortex. Vol. 4. Association and auditory cortices. A. Peters & E.G. Jones, (Eds.). London: Plenum Press, pp. 3-61
Pulvermueller, F. (1992) Constituents of a neurological theory of language. Concepts in Neuroscience 3:157-200.
Pulvermueller, F., Preissl, H., Eulitz, C., Pantev, C., Lutzenberger, W., Elbert, T. & Birbaumer, N. (1994a) Brain Rhythms, Cell Assemblies and Cognition: Evidence from the Processing of Words and Pseudowords. PSYCOLOQUY 5(48) brain-rhythms.1.pulvermueller.
Pulvermueller, F., Lutzenberger, W. & Birbaumer, N. (1994b) Electrocortical distinction of vocabulary types. Electroencephalography and Clinical Neurophysiology (in press).