Overview

1
Overview

• Why study eye movements?
• Peripheral oculomotor apparatus
• Five types of eye movements
• Anatomy, physiology and function

2
Why eye movements?

• Goal understand how the brain produces behavior
• Generate motor output based on current and past
  sensory input
• Eye movements are the right level of complexity
• (continuum from stretch reflex to playing squash)
• complex, yet highly constrained
• Other pluses
• Output is easy to measure
• Input is easily specified and delivered
• Peripheral apparatus is straightforward
• Good animal models available
• Why use animals for this?
• single neurons manipulate the system

3
Musculature
4
Muscle properties

• More complex than somatomotor muscle fibers
• 5 distinct fiber types (vs 2 - fast slow)
• Unclear why
• More proprioceptors (?)
• but proprioception is (too) slow
• Much higher innervation ratio (nerve
  endings/fiber)
• Built for speed, not for comfort
• 8 ms twitch time (2-3 times faster than fast
  somatomotor fibers)

5
Muscle innervation (oculomotor nerves)

• At rest, firing rate of an individual nerve is
  linear with eye position
• Different nerves have different slopes and
  offsets
• Sum to a non-linear increasing function that
  matches passive muscle properties

When the eyes move, you need still more force -
activity is proportional to position and velocity
(stay tuned)
6
Conjugacy

• Normally, the two eyes move as if yoked together
• Failure to yoke strabismus
• Strabismus may lead to reduced or absent vision
  in one eye amblyopia
• Bad cosmetic impair depth perception dont
  have a back-up eye. NIH goal (Hubel Wiesel)
  Lots of mechanisms to ensure that the eyes are
  aligned it seems that they must all fail in
  order to produce strabismus.
• Conjugacy is accomplished at the cranial nerve
  level

7
Conjugate eye movements
Oculomotor commands (conjugate)
Vergence commands
Interneurons
?
Oculomotor
Med rectus
8
Degrees of freedom

• d.f. - roughly, how many independent ways can the
  system be manipulated?
• Two eyes, 6 muscles per eye 12 d.f. (vs 30 in
  arm!)
• Muscles are paired (like flexor/extensor) 6 d.f.
• Conjugacy (horizontal, vertical and rotational)
  3 d.f.
• Vergence (horizontal) 4 d.f.
• How many do we need?
• Just 3!
• Are extra df's good or bad?
• Somatomotor avoid obstacles compensate for
  joints without 360 deg range of motion use one
  set of joints for power and another for
  precision etc. (Extrinsic versus intrinsic
  coordinates)
• Oculomotor system makes computation more
  difficult - degenerate solutions.
• (Has some use tortional VOR partial
  compensation for roll)

9
Methods of measuring

• Dodge, 1902 float mirror on the eye, bounce
  light off it onto a photographic plate on the
  wall
• Electro-oculogram (EOG) charge at back of retina
  moves when the eye moves (Wednesdays lab)
• Scleral search coil magnetic field induces a
  current in a coil, either surgically implanted in
  the eye (animals) or embedded within a contact
  lense (humans). Current is proportional to the
  angle of the coil
• Video-based methods A return to 1902, but now
  with computers

10
Five types of eye movements
VOR OKN Old systems, pre-date the fovea
11
Vestibulo-ocular reflex (VOR)

• Stabilize gaze by moving eyes opposite to the
  head
• Uses vestibular system for speed (Vision 35 ms
  to V1, 100 ms for movement, vs 10 ms for VOR)
• 3 neuron arc of Lorente de No 1930's (primary
  vestibular neuron, VOR interneuron, oculomotor
  neuron)
• Ratio of head to eye velocity gain. Ideally
  1 actually around 0.7
• Cares about velocity, not position (pre-fovea!),
  but there are fast phases to reset eyes
  position - keep it from pinning at the edge of
  the orbit ltnystagmusgt
• Open loop, but must be calibrated to be useful
• Good model system for learning plasticity
• Can be recalibrated swiftly and dramatically
• Good model system for short term memory
• velocity storage - extends time constant of
  vestibular periphery

12
Optokinetic reflex (OKN)

• Complements the VOR. Uses vision rather than
  vestibular input, so is sluggish but will
  maintain its response indefinitely (vs the
  transient VOR). (There is also a fast component
  to OKN.)
• Two pathways, sub-cortical (older, runs through
  brainstem accessory optic system and cortical
  (newer, runs through dorsal cortical pathway).
  Both converge on vestibular nucleus.
• Convergence on VN explains why OKN produces a
  motion (vestibular) sensation (waterfall
  illusion IMAX theatre waiting at traffic light).

13
Five types of eye movements
VOR OKN old systems, pre-date the fovea
Saccades pursuit new, built on VOR/OKN
14
Saccades

• Built on brainstem machinery for fast phases
• Purpose Rapidly realign the fovea
• Up to 800 deg/s!
• Cant see well, so do it fast
• Active suppression?
• Obligatory intersaccadic interval of 220 ms
• Why might this be?
• Spatial constancy - no sense of motion from
  saccades
• How does that arise?

15
Saccades

• Rapidly realign the fovea
• Usually not consciously monitored, so betrays
  (reveals) high order cognition (Yarbus)

16
Saccade control

• Open loop, driven by position error
• Too fast for feedback
• Need corrective saccades
• Need calibration, adaptation
• Stereotyped
• Consistent relations between amplitude, velocity
  and duration ('Main sequence')
• Experimentally useful characterizes true
  saccades (e.g., with stimulation)
• What is optimized/minimized?
• Acceleration, duration, force, jerk?
• VOR OKN retinal blur (velocity mismatch)
• Somatomotor what is minimized when you move your
  arm?

17
Saccade anatomy
Parietal
Frontal
Occipital
SEF
LIP
lobe
lobe
FEF
V1
SC
Brainstem circuitry
Cerebellum and basal ganglia
Oculomotor Neurons (CN III, IV, VI)
Muscles
18
Saccade-evoked activation (fMRI)
Look at the activity of single neurons
19
Superior colliculus - bursters
Flash a visual target for a saccade and record in
SC
How does this compare to responses in the cortex?
20
Cortical responses
In the cortex, responses are less brisk but
otherwise similar to those in SC. This is an
LIP neuron, but FEF, SEF and even V1 are grossly
similar. Sensory or motor? Use a paradigm that
separates out the visual stimulus from the motor
response.
21
Memory saccade paradigm
Separate sensory and motor events in time. We see
visual and motor activity in the same cell, which
blurs the distinction between sensory and motor
areas. (Similar effects in LIP, FEF and SC). The
transformation from sensory input to motor output
is gradual there are no sharp dividing
lines. This cell also shows memory activity.
22
Saccade anatomy
Parietal
Frontal
Occipital
SEF
LIP
lobe
lobe
FEF
V1
SC
Brainstem circuitry
Cerebellum and basal ganglia
Oculomotor Neurons (CN III, IV, VI)
Muscles
23
Smooth pursuit

• Behavior Track what we want to look at
• Built on machinery for slow phase VOR
• Control strategy
• Closed loop, driven by velocity error
• More than negative feedback it can predict!
• (well see this in lab)
• Anatomy
• V1 ? MT ? MST ? Pons ? Cbl (same areas as VOR) ?
  VN ? oculomotor nuclei
• There are also frontal pursuit areas recently
  discovered still unclear how they fit in.

24
Five types of eye movements
VOR OKN Saccades Smooth pursuit Vergence (fixatio
n)
25
Final common pathway

• All of these eye movement systems provide an eye
  movement command that specifies eye velocity -
  even the saccade system.
• Recall the eye muscles have passive properties
  which act like a spring and tend to center the
  eye.
• Therefore once the eyes have been moved out to
  some location, sustained force is required to
  keep them there. Thats why, when the eyes are
  at rest, firing rate is proportional to eye
  position.
• When the eyes move, force varies with both
  position and velocity.

26
Oculomotor firing rate is proportional to
position when the eyes arent moving
Firing rate
Eye position
27
When the eyes move, firing rate (force) varies
with both position and velocity
28
Final common pathway

• How is this accomplished? The brainstem final
  common pathway integrates velocity commands and
  then adds the integral to the command.
• Pulse-step Move the eye quickly and then hold it
  in place. Without the step, the eye would move
  to a new location but then slip back towards the
  center.
• A pulse-step mismatch would lead to either
  gaze-paretic nystagmus (gain lt 1) or unstable
  gaze (gain gt 1). This requires a perfect
  integrator - very challenging to build!
• Prepositus hypoglossi. Goldfish model system.

29
Oculomotor versus somatomotor systems