Extraocular motoneurons

Medial and lateral rectus muscles, which are innervated by motoneurons in the oculomotor and abducens nuclei, respectively, are primarily responsible for convergence and divergence. Figure 95.1 is a schematic diagram of these motoneurons, as well as some of the premotor inputs to them. Extraocular motoneurons show a burst-tonic pattern of activity for saccades and have a linear relationship between firing rate and eye position (Robinson, 1970) for steady fixation.

Figure 95.1..  

Schematic representation of horizontal oculomotor elements. Shaded ellipses indicate the oculomotor (OMN) and abducens (ABD) nuclei. The global layers of medial and lateral muscles are shown as inserting on the left (LE) and right eyes (RE). These layers are innervated by the subset of medial rectus (MR_G) and lateral rectus (LR_G) motoneurons. The orbital layers of these muscles are shown as inserting on muscle pulleys (P) and are innervated by the orbital subset of motoneurons (MR_O and LR_O). These MR_O and LR_O motoneurons appear to be responsible for moving the pulleys along an anterior-posterior axis (arrows). Abducens internuclear (AI) neurons are located in the abducens nucleus and send axons via the medial longitudinal fasciculus (MLF) to provide excitatory innervation to MR motoneurons and presumably oculomotor internuclear (OI) neurons as well. Many of the OI neurons are thought to project back to the ABD nucleus. Near response (NR) cells, which have a signal related to vergence and accommodation, appear to provide the needed vergence signal to MR motoneurons and probably provide the same signal to OI neurons. Two additional inputs (I₁ and I₂) are shown for ABD neurons. Hering's Law implies that I₁ and I₂ should be conjugate and vergence inputs, but recent work (Zhou and King, 1998) suggests that I₁ and I₂ are monocular ipsilateral and contralateral eye movement inputs.

Neurons in the Oculomotor Nucleus

Two studies (Gamlin and Mays, 1992; Mays and Porter, 1984) showed that nearly all putative medial rectus (MR*) motoneurons increased their activity for symmetrical convergence. This observation refuted earlier speculation (Alpern and Wolter, 1956) that a distinct subset of motoneurons was uniquely responsible for vergence movements. The comparison between the slope of the tonic firing rate-to-position relationship of a motoneuron for conjugate eye movements (versional gain) and that for purely symmetrical convergence movements (vergence gain) is of particular interest. Figure 95.2 illustrates this comparison for a sample of horizontal burst-tonic neurons in the medial rectus subdivisions of the oculomotor nucleus. Overall, the mean vergence gain of MR* motoneurons was essentially equal to that for changes in conjugate horizontal eye position, although the range of gains was much larger for convergence. Indeed, as Figure 95.2 shows, there is a very poor correlation between vergence gain and versional gain. If each motoneuron behaved in the same way for conjugate adduction as for convergence adduction, then all of the filled points in Figure 95.2 would fall along the line labeled “Ipsi eye.” About 10% of MR motoneurons show little or no change in activity for convergence (points near the “Version only” line in Fig. 95.2), and another 10% have extremely high gains for vergence when compared to their versional gains. These recording studies may well have included some oculomotor internuclear (OI) neurons, which are located within the medial rectus subdivisions of the oculomotor nucleus but are not motoneurons (Fig. 95.1). An analysis of OIs (Clendaniel and Mays, 1994), identified by antidromic stimulation from the contralateral abducens nucleus, revealed activity profiles for versional and vergence eye movements which largely overlapped those of the larger sample of MR* motoneurons (Fig. 95.2). One possible difference is that none of the OI neurons had the very high vergence gains seen in some MR* cells, but this difference may be due to the very small sample of identified OIs. Interestingly, the activity of the OIs matched the activity of MR* motoneurons for both version and vergence, even though the OIs projected to the contralateral abducens nucleus. If OIs conveyed useful information about vergence movements to the contralateral abducens, one would expect these cells to be aligned with the “Contra eye” line in Figure 95.2, which is not the case. This indicates that the OIs coordinate versional movements and not vergence.

Neurons in the Abducens Nucleus

The vast majority of abducens neurons decrease their activity for convergence (Gamlin et al., 1989a; Maxwell, 1991; Mays and Porter, 1984). Figure 95.3 shows the vergence and versional gains of abducens neurons and of horizontal burst-tonic fibers in the medial longitudinal fasciculus (MLF). The latter are presumed to be axons of abducens internuclear neurons (AIs), which have somata in the abducens nucleus and project, via the MLF, to provide excitatory input to medial rectus motoneurons (Steiger and Büttner-Ennever, 1979). The filled symbols in Figure 95.3 show data from abducens neurons that presumably represent both lateral rectus motoneurons and AIs. Open circles show data from antidromically identified AI neurons. Cells that exhibit the same behavior for versional and vergence abduction would fall along the “Ipsi eye” line, those that change their activity only for version would coincide with the “Version only” line, and those that fall along the “Contra eye” line would behave like contralateral MR motoneurons. Two important observations can be drawn from Figure 95.3. The first is that AIs (including MLF fibers) show the same pattern of responses as the overall population of abducens neurons, which presumably includes many lateral rectus (LR) motoneurons. The conclusion is that most LR motoneurons and AIs decrease their firing rate for convergence as well as for versional adduction. This is a critical observation since AIs provide powerful excitatory input to contralateral MR motoneurons. Since AIs decrease their activity for convergence, this means that they send an inappropriate signal to MR motoneurons during vergence movements. This implies that the AIs, like the OIs, coordinate conjugate horizontal eye movements and not vergence movements. The second observation is that, on average, abducens neurons do not reduce their firing rate as much for vergence adduction as for versional adduction. Indeed, most neurons in Figure 95.3 fall below the “Ipsi eye” line, which indicates equal firing rate change for version and vergence. This would suggest some degree of co-contraction of MR and LR muscles during convergence. However, recent measurements of LR and MR forces during convergence show no evidence of co-contraction (Miller et al., 2002). One possible reason for this apparent contradiction between muscle force measurements and motoneuron activity is that not all subsets of motor units develop equivalent forces for a given firing rate. Motoneurons responsible for larger decreases in muscle force during convergence may have been undersampled. Other explanations may involve the role of the trochlear motoneurons and muscle pulleys.

Neurons in the Trochlear Nucleus

Although LR and MR muscles are primarily responsible for vergence eye movements, there is a well-known excyclotorsion of the eyes associated with convergence (Allen and Carter, 1967). The magnitude of this torsion depends on elevation of the eyes as well as the vergence angle, with greater excyclotorsion seen with ocular depression. A study of trochlear motoneurons (Mays et al., 1991) showed a systematic decrease in activity associated with convergence that was far larger than that expected from the values associated with conjugate abduction of the innervated eye. Moreover, the magnitude of the decrease in firing rate increased with ocular depression. These findings are consistent with the behavioral observation of excyclotorsion with convergence and with the lateral tilt of Listing's plane during convergence (van Rijn and Van der Berg, 1993). The decreased superior oblique tension could aid adduction during convergence (Mays et al., 1991), but a recent analysis (Miller et al., 2002) of all muscle forces indicates that this cannot explain the discrepancy between the motoneuron and muscle force data.

Muscle Pulleys

Recent evidence has shown that the effective origin of some eye muscles is altered by passing through slings or pulleys associated with Tenon's capsule (Demer et al., 1995). These pulleys are shown schematically in Figure 95.1 as rings labeled “P.” It has been hypothesized that these pulleys provide a mechanical means for the eyes to obey Listing's Law. Indeed, it appears that the LR pulleys are moved posteriorly during convergence, while the MR pulleys shift anteriorly in such a way as to produce the lateral tilt of Listing's plane (Clark et al., 2000). Recent work (Demer et al., 2000) indicates that the orbital fiber layers of the LR and MR muscles insert on muscle pulleys and not on the globe itself. This implies that the activity of the motoneurons innervating orbital fibers in the pulley muscles (LR_O and MR_O in Fig. 95.1) may have very different activity patterns than the motoneurons innervating the global muscle fibers (LR_G and MR_G). Thus, the discrepancy between the LR motoneuron data and the absence of co-contraction may be due to recording a mixture of unidentified LR_O and LR_G motoneurons. To date, it has not been possible to distinguish between these motoneuron types with single-unit recording alone. The added biomechanical complexity of movable muscle pulleys makes it difficult to estimate muscle forces indirectly during convergence by observing the motoneuron firing rate from a combined motoneuron pool.