Robert Miller (1994) Cognitive Processing, but not Cell Assembly Ignition. Psycoloquy: 5(50) Brain Rhythms (2)

Volume: 5 (next, prev) Issue: 50 (next, prev) Article: 2 (next prev first) Alternate versions: ASCII Summary
PSYCOLOQUY (ISSN 1055-0143) is sponsored by the American Psychological Association (APA).
Psycoloquy 5(50): Cognitive Processing, but not Cell Assembly Ignition

Commentary on Pulvermueller et al. on Brain-Rhythms

Robert Miller
Department of Anatomy and Structural Biology
University of Otago Medical School
Dunedin, New Zealand


Pulvermueller et al.'s (1994) PSYCOLOQUY target article provides valuable data about the differential electrographic responses to word and pseudoword stimuli (found exclusively in the left hemisphere and presumed to represent a difference in cognitive processing). It is not clear, however, that this difference is an indication of cell assembly ignition; and even if it is, the difference would not be expected to apply selectively to high frequencies.


brain theory, cell assembly, cognition, event related potentials (ERP), electroencephalograph (EEG), gamma band, Hebb, language, lexical processing, magnetoencephalography (MEG), psychophysiology, periodicity, power spectral analysis, synchrony
1. Pulvermueller et al.'s (1994) target article reports interesting differences in the ERP (event-related potential) and 30 Hz electrographic activity with word and pseudo-word stimulus presentations. The differences do appear to be related to cognitive processing of words, rather than to their physical form, and, as might be expected, they emerge only in the hemisphere which has a predominant role in language. I have two major concerns, however, with the authors' claim that these differences reflect the selective activation of cell assemblies by word stimuli.

2. First, arguments are presented in paragraph 10 suggesting that oscillations mainly in the high frequency range should be characteristic of the activation of cell assemblies representing words. These arguments are based partly on anatomical data on axon calibres, from which the probable conduction time is inferred. These arguments are, I believe, invalid. Analyses of antidromic conduction in populations of single cortico-cortical axons typically find a wide range of conduction times. For example, in a study in the cat (Miller, 1975; with a conduction distance of about 10 mm), conduction times ranged from 2 to 44 msec. Similar data have been reported by Swadlow (1981; 1991; and others) in the rabbit. The morphological study cited by Pulvermueller et al. (Aboitiz et al., 1992) presents electron- microscopic data from the human corpus callosum. However, it is impossible to obtain such tissue from the human brain in a sufficiently fixed condition to preserve the inconspicuous but very numerous unmyelinated axons. The axon count in this study is likely to have been biassed in favour of the larger and more robust myelinated fibers.

3. In studies of cerebral white matter from nonhuman species, where perfect fixation is possible, about 50% of callosal axons are unmyelinated, with calibres ranging from about 0.1 um to 0.7 um (Fleischhauer and Wartenberg, 1967; Seggie and Berry, 1972; Berbel and Innocenti, 1988). These calibres probably correspond to conduction velocities of from 1 m/s to 0.1 m/s, much slower than the figure of 10 m/s assumed by the Pulvermueller et al. (which probably corresponds to the fastest-conducting myelinated fibers). Evoked potential studies are also cited by the authors to justify the assumption of relatively rapid conduction between cortical areas. However, this method is again biassed in favour of the more rapidly conducting axons in a population with a wide range of conduction velocities. In this case the bias arises because rapidly conducting axons will tend to transmit signals to their destination relatively synchronously and therefore there is greater possibility of temporal summation. Signals transmitted along slow-conducting axons would be dispersed in time, and hence would not contribute to the first part or the peak of an evoked potential, which is usually where latency measurements are made. Such a bias would not occur under natural stimulation, when there is no sudden synchronous activation of many neurons.

4. Assume that the conduction distance between Broca's and Wernicke's areas is about 100 mm. Taking into account the full spectrum of axon calibres and conduction velocities, conduction time could be anything from 10 msec to 1000 msec. This could give reverberation frequencies of anything from 1 to 100 Hz. Assuming that cortical electrographic rhythms arise from reverberation in cortico-cortical loops, and arguing just from anatomical data, one might expect that the activation of a transcortical cell assembly involving both Broca's and Wernicke's areas could include a wide mixture of many frequencies within this extensive range.

5. Second, it is not clear that the gamma band changes, which are differential to the word/pseudoword stimuli, are actually related to the SPECIFIC cognitive processing of each individual word (i.e., to its meaning). Although the differential responses are probably related to differences in semantic attributes of the stimuli, they could be related to less specific processes, such as the activation of word meaning in general. In principle, it might be possible to demonstrate a more specific role of changes in oscillatory activity in the EEG related to semantic attributes specific to each word. For example, the pattern of frequencies for which there is a difference between a word and matched pseudoword might be different for different words.

6. I have one other, smaller point to ask: in paragraph 3 the authors assume that they are dealing with cortico-cortical projections from motor to auditory systems; later they assume, more specifically, that conduction between Broca's and Wernicke's area is important. Apart from gross anatomical studies of dissectable fiber bundles, I have been unable to find any more precise evidence of this connection in experimental primates, let alone in humans. Does such evidence exist?

7. Overall, I am as yet unconvinced that these results indicate the ignition of specific cell assemblies related to word meanings. I suspect that they are indications of a different process of a more general nature. What this process is, I cannot guess at present (though I find the results obtained interesting and intriguing). This is part of a larger uncertainty (for me) as to whether gamma band rhythms arise from the biophysical properties of single neurones or from the network properties of very many interconnected neurones. If the latter proves to be the correct interpretation, there are a variety of more detailed network explanations to be considered, not just reverberation in cortico-cortical loops.


Aboitiz, F., Scheibel, A.B., Fisher, R.S. and Zaidel, E. (1992) Fiber composition of the human corpus callosum. Brain Research 598:143-153.

Berbel, P. and Innocenti, G.M. (1988) The development of the corpus callosum in cats: light and electronmicroscopic study. J. Comp. Neurol. 276, 132-156.

Fleischhauer, K. and Wartenberg, H. (1967) Electronenmikroscopische Untersuchungen uber das Wachstum der Nervenfasern und uber das Auftreten von Markscheiden in Corpus callosum der Katze. Zeitschrift fur Zellforsch. 83, 568-581.

Miller, R. (1975) Distribution and properties of commissural and other neurons in cat sensorimotor cortex. J. Comp. Neurol. 164, 361-374.

Pulvermueller, F., Preissl, H., Eulitz, C., Pantev, C., Lutzenberger, W., Elbert, T. and Birbaumer, N. (1994) Brain Rhythms, Cell Assemblies and Cognition: Evidence from the Processing of Words and Pseudowords. PSYCOLOQUY 5(48) brain-rhythms.1.pulvermueller.

Seggie, J. and Berry, M. (1972) Ontogeny of interhemispheric evoked potentials in the rat: significance of myelination of the corpus callosum. Exp. Neurol. 35, 215-232.

Swadlow, H.A. (1981) Efferent systems of the rabbit visual cortex: laminar distribution of the cells of origin, axonal conduction velocities and identification of axonal branches. J. Comp. Neurol. 203, 799-824.

Swadlow, H.A. (1991) Efferent neurons and suspected interneurons in second somatosensory cortex of awake rabbit: receptive fields and axonal properties. J. Neurophysiol. 66, 1392-1409.

Volume: 5 (next, prev) Issue: 50 (next, prev) Article: 2 (next prev first) Alternate versions: ASCII Summary