Modularity hypothesis

Poffenberger manipulated the strength of the input visual signal in the SRT paradigm by comparing monocular and binocular presentations. He noted that overall RT decreased with binocular input, but curiously, he did not analyze whether the CUD, that is, the IHTT, was affected as well. In 1971, Berlucchi and colleagues manipulated stimulus eccentricity, found that it did not affect the CUD, and concluded that interhemispheric transfer in the SRT paradigm is not visual. They reasoned that callosal fibers joining visual areas on the two sides are restricted to the neighborhood of the vertical meridian, so if transfer is visual, then more eccentric targets should take longer to transfer than less eccentric ones. Milner and Lines (1982) argued similarly (for acallosals, but the argument carries over to normals) that increased stimulus intensity increases the rate of neuronal recruitment in visual pathways, including in-terhemispheric ones.

More generally, one can articulate a modularity hypothesis according to which if interhemispheric transfer occurs through a particular callosal channel that interconnects specific cortical modules, then only the manipulation of a stimulus parameter that affects processing in those modules can affect the corresponding channel and thus interhemispheric transfer time in the same direction. Thus, only visual parameters can affect visual transfer, and visual parameters, such as intensity, eccentricity, or visual complexity, can affect only visual transfer. Similarly, cognitive parameters can affect cognitive transfer, and motor parameters can affect motor transfer. Thus, decision complexity should affect the transfer of cognitive/abstract information but not visual or motor transfer, whereas the complexity of the motor response pattern should affect motor transfer exclusively. Further, transfer along a particular visual pathway may be sensitive to one visual parameter (e.g., eccentricity) but not to another (e.g., intensity).

This view presupposes a decomposition of the information-processing sequence into several functionally independent components, including separate stages of visual analysis and motor programming. In this manner the visual and motor stages are modular (i.e., mutually impenetrable and encapsulated; Fodor, 1983). Thus, a change in visual parameters of the task may affect visual analysis but no subsequent stages. If the CUD is invariant to the manipulation of a visual parameter, however, this does not necessarily imply nonvisual transfer. Transfer may be between visual areas that are far enough along not to be sensitive to the parameters used. Therefore, the CUD may be sensitive to other untried or as yet undiscovered visual parameters.

However, it is possible that the visual and motor stages are not modular. Transmission times for nonvisual pathways might therefore also be affected, cascade fashion, by changes in stimulus strength. The CUD in acallosal and commissurotomized patients could similarly result from nonvisual interhemispheric transfer or from the use of ipsilateral motor connections that are less effective than their contralateral counterpart (see Kinsbourne and Fisher, 1971; Zaidel, 1983; Jeeves and Milner, 1987; Trope et al., 1987). Nevertheless, it may still be possible to determine whether transfer occurs along visual or nonvisual pathways by manipulating two, rather than one, visual stimulus parameters.

If stimulus intensity and eccentricity manipulations have equivalent effects on the CUD, then the findings may still be equivocal, depending on whether or not modularity holds. However, if intensity and eccentricity

Table 1

Summary of the effects of motor, attentional, and visual manipulations on the CUD

RT CUD

Motor

Iacoboni & Zaidel, spatial uncertainty — —

Iacoboni & Zaidel, RG between hems. + — Braun et al., '95

RG, redundancy gain manipulations can be shown to have disparate effects on the CUD, while having comparable effects on overall reaction times, then transfer is probably occurring along visual pathways that are more sensitive to one of the two parameters (assuming that the same pathway mediates transfer in both manipulations).

Table 1 summarizes the effects of visual, motor, and attentional parameters on the CUD in normal subjects in studies extending up to 1995. The informative cases are those in which the manipulation affects overall RT. The standard visual parameters are intensity and eccentricity, and they usually affect overall RT but not the CUD. St. John and colleagues (1987) did find an effect of eccentricity on the CUD, but the difference was small (e.g., a 0.9 ms longer CUD at 2° than at 15°), and the majority of subjects did not show it.

Attentional parameters include blocking of target VF, spatial uncertainty of target location, redundancy gain, and precuing of target location. Again, there are effects on overall latency but not on the CUD. Motor manipulations include bimanual responses, alternating fingers or alternating hand responses, and number of response presses. In contrast to visual and attentional parameters, motor manipulations do affect the CUD. This is consistent with conclusion that interhemispheric transfer in the SRT paradigm of Poffenberger is at a late, motor, information-processing stage.

The algebraic model of the CUD assumes strong modularity and might not be valid. However, as a strong theoretical position, it serves to clarify opposing accounts of the CUD and to focus disconfirming evidence.

Was this article helpful?

0 0

Post a comment