The interconnections between cerebral cortex and SC

Saccades are mediated by a system extending from the retina to the extraocular muscles that has been most extensively studied in the rhesus monkey, the animal model on which we will concentrate. The visuosaccadic system can be regarded as having three limbs, two of which are the afferent and efferent limbs (Fig. 98.1A). The afferent limb provides visual input from the retina. This pathway branches into two main routes, one coursing to the lateral geniculate nucleus and then to striate cortex (also known as V1) and another coursing to the superficial layers of the SC. The efferent limb provides motor output to the extraocular muscles. Motor commands are produced in the saccade generating circuitry in the brainstem, a collection of regions in the pons and midbrain that receives descending input from the SC and other structures.

Figure 98.1..  

The three limbs of the visuosaccadic system in the monkey. A, The afferent limb, which provides visual input from the retina, and the efferent limb, which causes muscle contractions. Structures with dashed outlines are subcortical. There are other targets of retinal projections, too, but they are beyond the scope of this chapter. SC, superior colliculus (s, superficial layer; i, intermediate layer); LGN, lateral geniculate nucleus; V1, striate cortex; SGC, saccade generating circuitry. B, The intermediary limb, which connects the afferent and efferent limbs. This limb is a network of brain regions; prominent among them are the frontal eye field (FEF) and the lateral intraparietal area (LIP).

The third limb is intermediary, connecting the afferent and efferent limbs (Fig. 98.1B). It is in this vast network of “in-between” areas that activity related to cognitive processes presumably resides, and interconnections between cerebral cortex and the SC are a major component of this network. Several regions of cerebral cortex project monosynaptically to the SC, and the SC reciprocates by projecting disynaptically up to many cerebral cortical regions via the thalamus (for reviews of the anatomy, see Leichnetz and Goldberg, 1988; Sparks and Hartwich-Young, 1989).

The SC and certain cortical regions are necessary for saccade generation. Reversible inactivation of the frontal eye field (Fig. 98.1B, FEF) causes severe deficits in the ability to make saccades to remembered targets (Chafee and Goldman-Rakic, 2000; Dias and Segraves, 1999; Sommer and Tehovnik, 1997), and similar effects have been reported for inactivation of the lateral intraparietal area (Fig. 98.1B, LIP) (Li et al., 1999). Reversible inactivation of the SC causes deficits in saccade production for all saccades made to either remembered or visual targets (Aizawa and Wurtz, 1998; Hikosaka and Wurtz, 1985; Lee et al., 1988; Schiller et al., 1987). Permanent ablation of either the SC or the frontal eye field causes serious deficits for a week or so, after which monkeys largely recover, but combined bilateral lesions of both the SC and the frontal eye field permanently devastate saccadic behavior (Schiller et al., 1980). The influence of many cortical regions seems to depend in large part on their projections to the SC; for example, the ability to evoke saccades electrically from the frontal eye field is severely impaired by inactivating the SC (Hanes and Wurtz, 2001), and the ability to evoke saccades electrically from the parietal and occipital lobes is abolished by ablating the SC and the frontal eye field (Keating and Gooley, 1988; Schiller, 1977).

In sum, there is good reason to suspect that the overall network composed of the cerebral cortex and the SC is important for saccade generation. In the first half of this chapter, we will review what is known about the signals descending from cerebral cortex to the SC. Investigated so far have been three direct projections emanating from the frontal eye field, the lateral intraparietal area, and the striate cortex (Fig. 98.2A) and one indirect pathway, a basal ganglia route that relays signals from cortex to the SC via the substantia nigra pars reticulata (Fig. 98.2B). Then we will discuss the signals flowing in ascending pathways from the SC to cerebral cortex. Two such pathways are starting to be well characterized, one that reaches the frontal cortex via the mediodorsal thalamus and another that reaches the parietal cortex via the pulvinar (Fig. 98.2C).

Figure 98.2..  

The descending projections and ascending pathways connecting cerebral cortex and the SC that have thus far been examined. A, Direct projections from cerebral cortex to the SC. The frontal eye field (FEF) and lateral intraparietal area (LIP) project to the SC intermediate layers, and the striate cortex (V1) projects to the SC superficial layers. B, Indirect route from cerebral cortex to the SC via the basal ganglia. Many regions of cerebral cortex project to the caudate nucleus (CN), which in turn projects to the substantia nigra. The substantia nigra pars reticulata (SNr) relays these signals to the intermediate SC. C, The ascending pathways from the SC to cerebral cortex. The intermediate SC projects to the mediodorsal thalamus (MD), which in turn projects to the frontal eye field and other prefrontal regions. The superficial SC projects to the pulvinar (Pulv.), which in turn projects to the lateral intraparietal area and other extrastriate regions.

To eavesdrop on the dialogue between cerebral cortex and the SC, investigators have identified neurons that talk back and forth between the areas using two methods: antidromic and orthodromic activation. Antidromic activation is used to see if a recorded neuron in one region (e.g., a part of cortex) projects to another region (e.g., the SC). Figure 98.3A shows the logic. If electrical stimulation in region B causes a recorded neuron in region A to fire at a fixed latency, and if other tests including one known as the collision test are successful (see the review by Lemon, 1984), then it is concluded that the neuron in region A is being activated through its own axon; therefore, the neuron projects to region B. Orthodromic activation is used to see if a recorded neuron in one region is receiving input from another region (Fig. 98.3B). If stimulation in region A causes a recorded neuron in region B to fire at a jittery latency and the collision test fails, then it is concluded that the neuron is being activated through synapses; therefore, this neuron receives input from region A. After a neuron's connections are identified with one or both of these methods, it is then analyzed to determine what signals it is carrying, as discussed next.

Figure 98.3..  

Methods for evaluating the signals sent between brain regions. A, In antidromic activation, one region, B, is stimulated while recording from a neuron in another region, A. Stimulation-evoked action potentials travel backward through the axon of the neuron and are recorded at the cell body. B, In orthodromic activation, one region, A, is stimulated while recording from a neuron in another region, B. Stimulation evoked action potentials travel forward through projections and drive the neuron via synapses, and resultant postsynaptic action potentials are recorded at the cell body. (Spike waveform pictures modified from Sommer and Wurtz, 1998.)