The smooth pursuit (SP) system is able to adapt to challenges associated with development or system drift to maintain pursuit accuracy. Short-term adaptation of SP can be produced experimentally using a step-ramp tracking paradigm with 2 steps of velocity (double-step paradigm). Previous studies have demonstrated that the macaque cerebellum plays an essential role in SP adaptation. However, it remains uncertain whether neuronal activity in afferent structures to the cerebellum shows changes associated with SP adaptation. Therefore, we focused on the dorsal–medial part of medial superior temporal cortex (MSTd), which is part of the cortico-ponto-cerebellar pathway thought to provide extraretinal signals needed for maintaining SP. We found that 54% of the SP-related neurons showed significant changes in the first 100 ms of response correlated with adaptive changes of initial pursuit. Our results indicate that some cortical neurons in MSTd could be inside the circuit involved in SP adaptation. Furthermore, our sample of MSTd neurons started their discharge on average 103 ms after SP onset. Therefore, we suggest that extraretinal signals carried in MSTd might be due to efference copy of pursuit eye velocity signals, which reflect plastic changes in the downstream motor output pathways (e.g., the cerebellum).
Smooth pursuit (SP) accuracy is maintained by adaptive mechanisms to keep the image of a moving object on the fovea. The SP system has the capability to adapt to changes associated with development and new behavioral demands (for review, see Leigh and Zee 2006). Learning new motor behaviors is aided by sensory error signals that occur in association with each attempted behavior. During SP, retinal error signals, defined as the difference between target and eye motion, play a role in maintaining pursuit accuracy. Previous studies have documented adaptive capability in the SP system using a step-ramp tracking task with a double step of velocity (double-step paradigm), where the target begins moving at one speed for first 100 ms and then changes to either a higher or a lower speed (Fukushima et al. 1996; Kahlon and Lisberger 1996; Takagi et al. 2000; Ono and Mustari 2007). This double-step paradigm is designed to introduce larger retinal error motion than occurs with single-velocity ramp tracking. These studies have shown significant adaptive changes in initial SP (first 100 ms) after 100–200 sequential trials. The first 100 ms of pursuit is thought to be an open-loop response that occurs before the time of the visual feedback. It is known that extrastriate cortex such as the middle temporal (MT) and medial superior temporal (MST) begin the process of converting visual motion into commands for SP (e.g., Newsome et al. 1985, 1988; Dürsteler and Wurtz 1988). These cortical areas, which provide sensory-motor–related signals for SP, could also be involved in visually guided motor learning.
Previous lesion and single-unit recording studies have demonstrated that the cerebellum, including the floccular complex and oculomotor vermis, play crucial roles in pursuit initiation and adapting SP (Kahlon and Lisberger 2000; Takagi et al. 2000). Furthermore, we recently reported that SP adaptation was severely impaired after muscimol inactivation of the dorsolateral pontine nucleus (DLPN) (Ono and Mustari 2007). The DLPN relays signals from cortical area MSTd to the dorsal/ventral paraflocculus and vermis of the cerebellum (Glickstein et al. 1980, 1994; Nagao et al. 1997; Distler et al. 2002). It is known that MSTd neurons carry extraretinal signals correlated with volitional SP (e.g., Newsome et al. 1988; Ilg and Thier 2003; Ono and Mustari 2006). These signals are thought to be delivered to the cerebellum over mossy fiber pathways leading to modulation of simple-spike activity. In fact, some Purkinje cells in floccular complex show significant changes in simple-spike and complex-spike firing associated with adaptation (Kahlon and Lisberger 2000). However, it has not been determined whether neuronal activity in structures supplying the cerebellum shows changes associated with SP adaptation. Therefore, our studies were directed at determining whether neurons in the cortical area MSTd have appropriate response dynamics to play a role in SP adaptation.
Materials and Methods
A detailed description of our surgical procedures can be found in recent publications (Mustari et al. 2001; Ono et al. 2005; Ono and Mustari 2009). Behavioral and single-unit data were collected from 2 normal juvenile rhesus monkeys (Macaca mulatta), weighing 5–8 kg. Briefly, surgical procedures were carried out under aseptic conditions in a dedicated surgical suite using isoflurane anesthesia (1.25–2.5%) to stereotaxically implant a head-stabilization post and recording chamber (Crist Instruments, MD). In the same surgery, a scleral search coil for measuring eye movements (Fuchs and Robinson 1966) was implanted underneath the conjunctiva of one eye using the technique of Judge et al. (1980). All surgical procedures were performed in strict compliance with NIH guidelines, and the protocols were reviewed and approved by the Institutional Animal Care and Use Committee at the University of Washington.
During all experiments, monkeys were seated in a primate chair (Crist Instruments) with their head stabilized in the horizontal stereotaxic plane. Visual stimuli were rear projected on a tangent screen 57 cm distant. All of our monkeys were extensively trained to perform a fixation task and track a small diameter (0.2°) target spot (produced by a laser light emitting diode moving in sinusoidal or step-ramp trajectories). Motion of the target spot was produced by a computer controlled 2-axis mirror galvanometer (General Scanning, Watertown, MA). Neurons in the MSTd were first classified as either large-field visual motion sensitive or SP-related neurons. Neurons that responded only during large-field (70° × 70°) visual motion, while the monkey fixated a centrally located stationary spot were not included in this study. Neurons that responded during SP of a small diameter (0.2°) target spot moving at low frequency (0.1–0.75 Hz; ±10°) were classified as SP-related neurons and were included in this study. SP neurons were tested while the monkey tracked a target that moved in 1 of 8 cardinal directions separated by 45°. All neurons were tested as monkeys tracked a target spot that moved with a step-ramp trajectory in a best direction. The size of the target step was adjusted so that the monkey initiated SP eye movements without early saccadic intrusions (Rashbass 1961). Usually the size of the step was between 2° and 4°. Neurons that evinced sensitivity to step-ramp tracking were studied during an adaptation paradigm. Adaptive changes of SP were produced as the monkey tracked double steps of target speed (double-step paradigm) that step-up (10–30°/s) or step-down (25–5°/s). In the step-up paradigm, the target begins moving at 10°/s for the first 100 ms and then changes to 30°/s for the remainder of the trial. In the step-down paradigm, the target begins moving at 25°/s for first 100 ms and then changes to 5°/s for the remainder of the trial. SP adaptation was evaluated during 100–200 trials for each adaptation paradigm. Control trials employing single-speed step-ramp tracking (ramp speed = 10°/s for step-up and 25°/s for step-down paradigm) were used before and after adaptation paradigms. We conducted one set of adaptation trials (either step-up or step-down) and associated control testing in a given experimental session. This approach ensured that animal behavior would be optimal during the single-unit recording.
Data Collection and Analysis
Eye movements were detected using electromagnetic methods (Fuchs and Robinson 1966) and precision hardware (CNC Electronics, Seattle, WA). Eye and target position feedback signals were processed with antialiasing filters at 400 Hz using 6-pole Bessel filters. Those signals were digitized at 1 kHz with 16-bit precision using CED-Power1401 hardware (Cambridge Electronic Designs, Cambridge, England). Single-unit activity was recorded from neurons in MSTd using customized epoxy-coated tungsten microelectrodes (Frederick-Haer Corporation, Brunswick, ME). Single-unit action potentials were detected with either a hardware window discriminator (Bak Electronics, Mount Airy, MD) or a template matching algorithm (Alpha-Omega, Israel) and represented by a TTL pulse which was registered at high precision as an event mark in our data acquisition system (CED Power1401). During analysis, neuronal response was represented as a spike density function that was generated by convolving spike times with a 5-ms Gaussian function (Richmond et al. 1987).
Model Fitting and Optimization
We attempted to reconstruct the individual response profiles of SP-related neurons in MSTd by using combinations of position (E), velocity (E′), and acceleration (E′′) of eye motion. Velocity and acceleration data were filtered using an 80-point finite impulse response digital filter with a passband of 50 Hz. The spike density function was also filtered at 50 Hz to reduce the variability in the unit response. Saccades were marked with a cursor on eye velocity traces and were removed. After desaccading, the missing eye data were replaced with a linear fit connecting the pre- and postsaccadic regions of data using custom Matlab routines (Mathworks, Natick, MA). Averaged data, taken from at least 10 trials in which the animal performed SP, were then used to identify coefficients in the following model:
Localization of the MSTd
We verified that our neurons were located in MSTd by functional criteria (e.g., response continues during a target blink and response latencies with respect to pursuit onset) and its stereotaxic location. Recording chambers were stereotaxically implanted with their centers located at posterior = 5 mm, lateral = 15 mm. We further verified that our recording locations were in MSTd by using magnetic resonance imaging (T1-weighted fast spin-echo; Siemens, 3-T magnet) and electrode track depth measurements taken from microdrive readings while recording SP neurons in MSTd. In Figure 1, we show a representative MRI image in the coronal plane (Fig. 1A) at a site located 5 mm posterior to ear bar zero, which provides further confirmation that our recording sites were well placed in the MSTd. Furthermore, the response properties of our MSTd pursuit neurons are consistent with those reported in previous single-unit recording studies with histological verification (Ono and Mustari 2006).
Response Properties of MSTd Neurons during Step-Ramp Tracking
We isolated 130 units in 2 monkeys for this study. We recorded 35 SP-related neurons in the MSTd and 95 neurons that responded only to large-field visual motion (monkey H, n = 24, SP, n = 57, visual motion; monkey T, n = 11, SP, n = 38, visual motion). For the present study, only SP-related neurons (n = 35) were included. The majority (69%) of neurons encountered in the MSTd area responded only to large-field visual motion. Figure 1B illustrates the response of a typical MSTd neuron during step-ramp SP including a target blink condition. This neuron discharged maximally during rightward SP. The neuronal discharge starts on average 130 ms after pursuit onset. To test for the presence of an extraretinal signal, we extinguished the target spot for a period of 150 ms during steady-state pursuit (Fig. 1B, top trace). Although eye velocity showed a small reduction during a target blink, the pursuit-related neuron did not show a clear drop in modulation. In our sample of MSTd SP neurons, neuronal discharge continued during target blinks. The mean value of blink/nonblink response ratio was 94.2 ± 11.5% (n = 35) and the range of these response ratio was 84.5–125.6% (median 93.2%). These results were consistent with those described in previous studies from our laboratory and in other reports (e.g., Newsome et al. 1988; Ono and Mustari 2006; Ono et al. 2010). Figure 1C shows latencies of the MSTd neurons with respect to pursuit onset. All of our MSTd SP neurons had response onsets following pursuit initiation (mean: 103 ± 41 ms).
Adaptation Alters Neuronal Discharge in MSTd Neurons
Figure 2 illustrates the time course of behavioral responses during the step-up adaptation paradigm and the concomitant change in firing rate of a representative MSTd neuron. In this paradigm, the target begins moving at 10°/s for first 100 ms and then changes to 30°/s for the remainder of the trial. The average eye acceleration in first 100 ms of SP during adaptation is plotted as a function of adaptation trial number (Fig. 2, top panel). Control trials employing single-speed step-ramp tracking (ramp speed = 10°/s) were used before and after adaptation to estimate the magnitude of adaptation. Bottom panels show that the firing rate of a single-MSTd neuron changes with adaptation (pre-, early-, late-, and postadaptation). All of the neurons reported here were recorded continuously during 30–40 min required for at least 100–150 double-step trials and 30–40 control trials. As the initial eye acceleration shows significant adaptive changes after 100 trials during adaptation (97.3 ± 18.8°/s2, first 10 trials; 194.5 ± 22.6°/s2, last 10 trials; P < 0.001, unpaired t-test), the average firing rate in the first 100 ms significantly increased (P < 0.001, unpaired t-test) from 50.5 ± 16.8 spikes/s (first 10 trials) to 78.5 ± 18.1 spikes/s (last 10 trials).
Figure 3 shows comparisons of eye velocity traces and MSTd neuronal responses pre- and postadaptation during single-speed control testing. After step-up adaptation trials (over 100 trials), control testing at 10°/s shows that eye velocity overshoots the target, revealing an adapted initial SP response (Fig. 3A,B). Figure 3A (cell # 6) illustrates a representative MSTd neuron that showed a significant increase in firing rate following adaptation (preadaptation = 28.6 ± 7.2 spikes/s; postadaptation = 63.3 ± 25.2 spikes/s; P < 0.001, unpaired t-test) correlated with adaptive change in initial eye acceleration (preadaptation = 68.8 ± 21.5°/s2; postadaptation = 119 ± 40.7°/s2; P < 0.01, unpaired t-test). In contrast, the neuron shown in Figure 3B (cell # 2) had similar responses across adaptation epochs (preadaptation = 18.3 ± 6.5 spikes/s; postadaptation = 19.2 ± 7.3 spikes/s; P = 0.77, unpaired t-test), even though initial eye acceleration shows significant adaptation (preadaptation = 65.6 ± 17.9°/s2; postadaptation = 101.8 ± 31.4°/s2; P < 0.01, unpaired t-test). In contrast, after step-down adaptation trials, control testing at 25°/s shows that eye velocity decreased postadaptation (Fig. 3C,D). Figure 3C (cell # 3) illustrates an MSTd neuron that showed a significant decrease in firing rate following step-down trials (preadaptation = 52.5 ± 19.4 spikes/s; postadaptation = 32.7 ± 7.9 spikes/s; P < 0.01, unpaired t-test) correlated with adaptive change in initial eye acceleration (preadaptation = 170.6 ± 57.2°/s2; postadaptation = 105.2 ± 33.9°/s2; P < 0.01, unpaired t-test). In contrast, the neuron shown in Figure 3D (cell # 4) had similar responses across adaptation epochs (preadaptation = 55.0 ± 17.6 spikes/s; postadaptation = 53.4 ± 12.2 spikes/s; P = 0.82, unpaired t-test), even though initial eye acceleration shows significant adaptation (preadaptation = 153.6 ± 41.6°/s2; postadaptation = 107.1 ± 28.9°/s2; P < 0.05, unpaired t-test).
Figure 4A shows changes in average firing rate during the first 100 ms of SP plotted as a function of the concomitant change in initial eye acceleration (first 100 ms) pre- and postadaptation. Filled circles indicate neurons with significant changes (P < 0.05, unpaired t-test) in firing rate postadaptation. Of 35 neurons, 19 neurons showed significant increases or decreases in firing rate postadaptation, while 16 neurons did not show significant (P > 0.05, unpaired t-test) changes in firing rate postadaptation. Furthermore, we applied regression analysis to identify correlations between initial eye acceleration and initial firing rate in 35 neurons which were tested during step-up or step-down adaptation paradigms. The fit obtained using linear regression had a coefficient of 0.59 (n = 35, P < 0.001), indicating that the firing rate tends to increase during step-up adaptation and decrease during step-down adaptation. Firing rates of each neuron as a function of target speed in step-ramp testing are shown in Figure 4B. Adapting (Fig. 4B-a) and nonadapting (Fig. 4B-b) MSTd neurons had similar nonsaturating responses during different speeds of control testing. Therefore, absence of altered neuronal response during adaption was not an artifact of potential saturation effects.
Model Testing Pre- and Postadaptation
We further examined the response properties of our MSTd neurons by applying a modeling procedure employing multiple linear regression. The model estimation procedure for the unit illustrated in Figure 2 is shown in Figure 5A (preadaptation) and B (postadaptation). Panels a–c in Figure 5 illustrate the eye motion components (position, velocity, and acceleration) that were used to make up the models. Panel d illustrates the contribution of each term of the model toward the total fit. Panel e illustrates the experimentally derived unit spike density function and the corresponding model estimated fit. The fits obtained using this 3-component model pre- and postadaptation had CDs of 0.86 and 0.82, respectively. Examination of each component of this model (panel d) indicates that eye velocity contributes most to the unit response during step-ramp tracking, while contributions from position and acceleration are relatively small.
Figure 6 illustrates that a 3-component model, which uses eye motion parameters, provided a good fit to the experimentally derived data in adapted neurons (filled circles) pre- (CD = 0.82 ± 0.10; n = 19) and postadaptation (CD = 0.81 ± 0.09; n = 19) and nonadapted neurons (open circles) pre- (CD = 0.80 ± 0.09; n = 16) and postadaptation (CD = 0.78 ± 0.08; n = 16). The model fits did not show significant differences between pre- and postadaptation for adapted neurons (P = 0.84, paired t-test) and nonadapted neurons (P = 0.63, paired t-test), which indicate that MSTd pursuit responses are highly dependent on eye motion parameters pre- and postadaptation.
Sensitivity of MSTd Neurons Pre- and Postadaptation
Figure 7 shows sensitivities of MSTd neurons pre- and postadaptation obtained during step-up (Fig. 7A) and step-down (Fig. 7B) paradigms. The sensitivities (spikes/s/°/s) in initial responses (first 100 ms) of adapted neurons (filled circles) did not show significant differences between pre- and postadaptation during the step-up paradigm (11.4 ± 5.5, pre; 12.3 ± 5.3, post; P = 0.66, paired t-test) or step-down paradigm (7.2 ± 2.0, pre; 7.0 ± 2.6, post; P = 0.89, paired t-test). In contrast, all the data points of nonadapted neurons (open circles) were either below (Fig. 7A) or above (Fig. 7B), the equality line drawn on the plot, showing that the neuronal sensitivities tend to decrease during step-up paradigm (8.3 ± 2.8, pre; 6.3 ± 2.4, post) and increase during step-down paradigm (6.5 ± 3.6, pre; 8.5 ± 4.3, post). However, these differences between pre- and postadaptation sensitivities were not significantly different (P > 0.05, paired t-test).
The present study was designed to consider whether MSTd SP signals were involved in adaptive changes of initial pursuit, revealed as overshooting or undershooting eye motion during control testing. We found that half our population of SP neurons showed significant changes in firing rate associated with adaptation. However, it is still unclear how different types of cortical neurons influence cerebellar signal processing for ongoing SP and learning. It is important to determine whether neurons in MSTd have responses consistent with producing adaptation or whether their activity simply reflects downstream changes after learning. Here, we discuss the implication of our results and whether MSTd neurons could contribute to adaptive changes of initial SP.
Existence of an Extraretinal Signal in MSTd
We sampled MSTd neurons with strong modulation during SP. These MSTd neuronal responses lagged SP onset, often by more than would be expected based on a typical visual latency related to retinal image motion. Furthermore, all of these neurons continued their discharge during target blinks, as we reported in earlier studies (Ono and Mustari 2006; Ono et al. 2010). In contrast, neurons in the nucleus of the optic tract (NOT) (Mustari and Fuchs 1990; Das et al. 2001), the MT area, or lateral–anterior region of MST (MSTl) (e.g., Newsome et al. 1988) with parafoveal visual sensitivity, respond at relative short latency (∼50 to 80 ms) following target motion onset. These neurons show a significant decrement (e.g., blink/nonblink ratio < 50%) in firing rate during target blink (Newsome et al. 1988; Mustari and Fuchs 1990; Das et al. 2001). Therefore, these results indicate that MSTd pursuit neurons might carry directional extraretinal signals independent of continuing visual inputs (Newsome et al. 1988; Ilg and Thier 2003; Ono and Mustari 2006; Ono et al. 2010).
Previous studies also have demonstrated that MSTd neuronal discharges often follow the onset of the SP eye movements (Newsome et al. 1988; Ilg and Thier 2003; Akao et al. 2005; Ono et al. 2010). In recent studies, we have shown that MSTd SP neurons were not modulated during the vestibular ocular reflex in darkness, which argues against a proprioceptive mechanism for their extraretinal signals (Ono and Mustari 2006). Thus, we suggest that the late onset of MSTd neuronal response could be an efference copy of SP commands that arrives through feedback circuits. Although a possible feedback circuit, involving thalamocortical pathways, has been partially identified (e.g., Tanaka 2005), we still do not know the neural mechanisms responsible for the long latencies of response with respect to pursuit onset.
Figure 8 diagrams pathways, which could be involved in SP and its adaptation. At least some MSTd neurons may receive an efference copy of pursuit command signals through feedback circuits (see below). Such a signal could provide maintained drive for SP as eye velocity approaches target velocity and visual error–driven neurons show less firing. Our studies did not address this issue per se. In contrast, parafoveal visual neurons in NOT and MT start firing before pursuit onset and their response declines once eye velocity reaches the target during step-ramp tracking (e.g., Newsome et al. 1988; Das et al. 2001;,Mustari et al. 2009). These results support the suggestion that initial pursuit (open-loop period) is driven strongly by retinal motion signals from parafoveal neurons. As pursuit eye velocity approaches target velocity, extraretinal signals carried by MSTd neurons could play a role in maintaining SP.
Role of MSTd Extraretinal Signals in Smooth Pursuit Adaptation
We used a double-step paradigm to produce adaptive changes in SP initiation. For example, following step-up adaptation trials (over 100 times repetition), single-velocity control testing at 10°/s shows that eye velocity overshoots the target (Figs 2 and 3), revealing an adapted initial SP response. This overshooting eye motion could be due to plastic changes of pursuit-related neuronal responses, which could occur in cortico-ponto-cerebellar and cortico-pretectal (accessory optic system) pathways. If adaptation occurs in downstream structures (e.g., cerebellum), we might not expect to observe changes in neuronal activity in cortical areas. For example, the frontal eye field (FEF) pursuit area contains neurons with responses related to SP initiation (Gottlieb et al. 1994; Tanaka and Fukushima 1998; Lisberger 2010) and these neurons project to the rNRTP (Brodal 1980; Ono and Mustari 2009), which targets the oculomotor vermis (Brodal 1982). However, the neuronal response of most FEF SP neurons did not change following SP adaptation using a double-step paradigm (Chou and Lisberger 2004). These investigators suggested that adaptation occurs downstream from FEF. Consistent with this suggestion, Takagi et al. (2000) found that lesions of the vermis impaired SP initiation and adaptation. Furthermore, at least some Purkinje cells in the floccular complex show significant changes in simple-spike response following SP adaptation (Kahlon and Lisberger 2000).
In this study, we have found that some MSTd neurons showed significant changes in initial firing rate following adaptation. Note that the eye velocity sensitivities (spikes/s/°/s) of these adapted neurons were unchanged between pre- and postadaptation. Furthermore, using multiple linear regression modeling, we showed that MSTd pursuit responses are highly dependent on eye velocity signals pre- and postadaptation. These results indicate that MSTd neurons could carry signals involved in adapted eye motion of overshooting or undershooting initial pursuit. However, the common finding that MSTd neuronal response follows SP onset argues against a role in driving the adapted initial pursuit response. Taken together, our findings support the suggestion that the changes in MSTd neuronal response that we observed are due to efference copy signals reflecting plastic changes in downstream motor output pathways (e.g., cerebellum).
We have also found that some MSTd neurons in this study did not change in initial firing rate following adaptation. Similar findings have been shown in FEF pursuit neurons (Chou and Lisberger 2004) and some simple-spike responses of Purkinje cells (Kahlon and Lisberger 2000). It has also been demonstrated that these simple-spike responses are related to pursuit or gaze eye velocity (Lisberger and Fuchs 1978). Since neuronal responses in our MSTd neurons are highly correlated with pursuit eye velocity, nonadapted neurons might play a role in maintaining steady-state eye velocity during SP through feedback circuits.
A delayed feedback signal from the cerebellum offers one possible explanation for the late onset of MSTd neuronal discharge relative to pursuit onset. However, the latencies (103 ms) of MSTd pursuit responses seem much longer than the conduction times required for a signal traveling from cerebellum/brainstem to cortical areas. An alternative hypothesis is that the delayed MSTd response is generated in the cortex. Such a driving signal is necessary for pursuit maintenance because the visual error signals in other pursuit areas (e.g., NOT and MT) decline as eye velocity matches target velocity. Further studies will be necessary to test competing hypotheses.
Possible Cortical Pathways Involving Smooth Pursuit Adaptation
Visual motion signals are processed in cortical areas MT, MST, and to some extent in FEF to produce partially formed commands for SP. These cortical signals must be processed further in the oculomotor regions of the cerebellum (for review, see Lisberger 2010). Extraretinal signals in MSTd neurons could originate as efference copy or corollary discharge of pursuit commands relayed through feedback pathways (Fig. 8). We know that corollary discharge signals for saccades reach the cerebral cortex through thalamocortical pathways (for review, see Sommer and Wurtz 2008). Although we lack direct evidence of a corollary discharge pathway for pursuit, recent studies have provided evidence that thalamic neurons in the central thalamus are involved in the control of SP (Tanaka 2005). Thalamic neurons had a wide range of latencies, ranging from coincident to lagging pursuit onset. The source of SP signals in the oculomotor thalamus includes the vestibular, fastigial, and dentate nuclei (Lang et al. 1979; Kalil 1981; Asanuma et al. 1983; Lynch et al. 1994). The oculomotor thalamus is known to provide extraretinal signals to the FEF and the supplementary eye field. The FEF has reciprocal connections with MST (for review, see Lynch and Tian 2006), but we do not know whether extraretinal signals for SP are present in this projection.
Other cortical pathways including ones carrying visual motion information could also play a role in adaptation of pursuit initiation. It has been demonstrated that MT neurons with foveal/parafoveal visual receptive fields were modulated during SP (e.g., Komatsu and Wurtz 1988; Newsome et al. 1988). A recent study that coupled electrical stimulation of MT with SP indicates that MT could provide instructive signals appropriate for guiding SP adaptation (Carey et al. 2005). Similarly, we have revealed that electrical stimulation of the NOT (delivered at the usual time of the second ramp in target speed) could substitute for actual visual error information to produce SP adaptation (Ono and Mustari 2010). Anatomical and functional connectivity studies have demonstrated that MT sends strong inputs to the DLPN and NOT, which in turn project to the floccular complex of the cerebellum via mossy and climbing fiber, respectively (e.g., Hoffmann et al. 1992; Distler et al. 2002; for review, see Gamlin 2006). Therefore, our findings and those of other laboratories indicate that at least some cortical pursuit areas participate in circuits involved in SP learning.
National Institutes of Health Grants, NEI, EY019266 (S.O.) and EY013308 (M.J.M.); RR00166.
We thank Dr Ulrich Büttner for helpful discussions during the course of this study. Conflict of Interest : None declared.