The visual pathway from retina through the lateral geniculate nucleus to visual cortex in the cat is comprised of several parallel neuronal streams that independently analyze different aspects of the visual scene. The best known of these are the X and Y pathways that relay through the geniculate A laminae. Recent receptive-field studies of retinal and geniculate neurons suggest that there is a further elaboration of cell types at the level of the lateral geniculate nucleus. That is, two types of geniculate X cells with different temporal patterns of responses to visual stimuli are recognized, one with "nonlagged" features, exhibiting shorter response latencies and another with "lagged" features; all retinal X cells are nonlagged. We asked whether nonlagged and lagged responses represent different cell classes or two response modes of the same cells, perhaps under the control of nonretinal afferents to these relay cells. Accordingly, we studied the effects on appropriate receptive-field properties of electrical activation of the midbrain parabrachial region, which is a major nonretinal input to relay cells. Such parabrachial stimulation made each of the eight lagged X cells much more like nonlagged cells, and this stimulation completely transformed the lagged response profiles of six of the eight cells to nonlagged. We thus conclude that the property of lagged responsiveness, which is an emergent property of the lateral geniculate nucleus, is a different response mode of the same cells that can also display nonlagged responses, rather than representing different cell classes; furthermore, this switching between response modes is, at least partly, under the control of afferents from the parabrachial region. (1,3,4). These nonretinal inputs derive from local, y-aminobutyric acid-releasing (GABAergic and thus probably inhibitory) neurons, from the visual cortex, and from various brainstem sites (1-4). Probably chief among the brainstem afferents are cholinergic axons emanating from the parabrachial region (1, 2).A question still to be addressed is precisely what transformations of retinogeniculate transmission are controlled by these nonretinal afferents. Perhaps the most dramatic recent example of such a transformation stems from evidence that the two main parallel pathways, X and Y (7-9), that innervate the lateral geniculate nucleus from retina are transformed into at least three pathways to cortex. This is because one of them, the X pathway, seems to divide into "nonlagged" and "lagged" classes (10-13). All afferent retinal X axons are of the nonlagged type, which implies that lagged X cells are an emergent property of geniculate circuitry (10,11 Mfl at 100 Hz). We inserted one pair of bipolar stimulating electrodes to straddle the optic chiasm and introduced another into the midbrain parabrachial region (cf. ref. 16). We applied single pulses (50-to 100-,usec duration, 150-700 ,uA) across the chiasm electrodes to activate geniculate cells orthodromically, and we stimulated the parabrachial region wit...