115 Midterm II Pavlovian Eye-Blink Conditioning & Cerebellar Motor Learning
Inferior Olive (IO) passes three tests that show ___________ Describe the 3 tests
IO allows *US to control plasticity* in the interpositus nucleus 1. *Correlation* = Neurons in IO respond to US (puff) - Activity is correlated w the puff 2. *Necessity* = When IO is lesioned, CR functions as if US is absent - Without synapse, pairing CS and US will never produce eye-blink CR 3. *Sufficiency* = Electrical stim of IO can substitute for US; if CS is paired w artificial IO stimulation, eyeblink CR develops even though eyelid is not shocked/puffed! - Stimulating neuron in conjunction w CS tone is sufficient to produce eye-blink CR
Does the animal's' behavior affect the relationship bt CS and US?
No! It does not affect relationship in any way... this is determined *entirely* by experimenter It is neutral PRIOR pairing with CS, but then becomes a PREDICTOR of CS - The animal is NOT learning *to do* anything, it's learning to *expect something* - Cannot make air puff not happen when you hear the bell
What must happen for NMDA receptors to allow Ca2+ influx?
Need the TWO following things at same time: 1. *Depolarization* = Mg2+ block repelled out of channel 2. *Glutamate binding* = released from pre-synaptic terminal
Which neurons respond to tone in experimental and control groups
Neurons upstream of interpositus - Pontine nucleus - Cochlear nucleus = auditory neurons that respond purely to tone *response remains the same (does not change) after pairing CS+US*
When do interpositus neurons respond to tone?
Not before it's paired with US, but after it's paired with US
Interpositus nucleus
One of the *DEEP cerebellar* nuclei!
What messages does interpositus receive when CS turns on?
Receives two messages: 1. Input from pontine to blink blink blink - puff is coming 2. Inhibitory input from cerebellar purkinje cells to wait
Graph of Unconditioned Eyeblink Reflex
US (shock/airpuff) delivered -> Eyelid position closes almost immediately (at same time)
Pretraining (Baseline) stage of Pavlovian Conditioning
US elicits an *unconditioned response* (UR) from subject, but the CS elicits NO response Puff elicits a blink, but the tone does not
Control condition of interpositus neuron recording
Unpaired presentations of CS and US (did not occur together) do not cause eye blink CR Also, *interpositus* nucleus does not fire to the CS - no conditioned neural responses observed in control
What controls the timing of CR
*Cerebellar Cortex*! McCormick & Thompson: - Lesions of *anisoform + paramedian lobules* (these talk to eyelid circuit) of cerebellum -> *disrupt* timing of response -> W/o these portions of cerebellar cortex = response is executed *immediately after CS onset* (tone) and - animal still learns to blink -> but closes eye *too early* and reopens eyes before puff comes CONCLUSION: cerebellar cortex seems to time + "sculpt" response
What type of reflex is the eyeblink reflex?
*Cranial* reflex! Not a spinal reflex because the muscles involved in control occur in the face Very simple behavior, mechanism is similar to the withdrawal reflex (Principles are pretty much the same)
How does cerebellum know when it is the right time to blink?
*Long-term depression* (LTD) at Purkinje synapse: CS paired w US -> Climbing fibers (from IO) project to interpositus (strengthen mossy fiber inputs) AND to cerebellar cortex - Active pontine (CS tone) activate granule cells Simultaneous activation of *granule cell + climbing fiber onto purkinje* = WEAKEN granule cell's synapse onto purkinje cell Consequently, very moment puff occurs -> Purkinje cells *lose ability to block CR* specifically *at time when US occurs* and NOT at other times! - Allows Purkinje cells to precisely control timing of response TIMING is IMPORTANT FOR MOTOR COORDINATION!
What connects pontine nucleus + interpositus nucleus
*Mossy fibers* project from pontine -> interpositus = axons
Receptors that mediate synaptic plasticity
*NMDA* = glutamate's ionotropic receptor
In pavlovian conditioning, a _______ CS is paired with an __________ US
*Neutral* conditioned stim with a *appetitive or aversive* unconditioned stim neutral tone (neutral prior to being paired w CS) neutral bell (neutral prior to being paired w CS) aversive airpuff appetitive food (ring bell -> food for dog)
IO necessity test
- *Pre-training* lesions of IO completely *prevent* acquisition of eyeblink *CR* (but do NOT interfere w eyeblink UR) -> Animal will never learn conditioned response, but UR still occurs - *Post-training* lesions cause *extinction of CR* during subsequent CS-US pairings, as if US was not being delivered anymore - Interpositus doesn't know puff is happening anymore (because IO cannot deliver info about puff to it)
Training (CS-US pairing) phase
- CS+US paired *during* training: 1A. CS causes *pontine nucleus* to fire -> release glutamate onto *WEAK* interpositus synapses 1B. At *same time* US causes *IO* to fire -> release glutamate onto *STRONG* interpositus synapses 2. Strong US-evoked depolarization of interpositus by climbing fiber -> kicks *Mg block off* NMDA receptor at mossy fiber -> NMDA bound to CS-evoked glutamate from pontine 3. *NMDA* receptor can now pass current (on mossy fiber) -> allows *Calcium* to enter interpositus neuron 4. what does calcium do?
Difference between copies of cerebellum circuit involved in conditioning
- Circuit = repeated millions of times over in cerebellum Each copy of of circuit receives *different "CS inputs"* - proprioceptive, vestibular, auditory, visual, etc AND sends *different "CR otuputs"* - fingers, legs, arms, neck, eyes, etc In concert w one another, these circuits make it possible for ppl to learn skills that require *complex motor coordination* = walking upright, athleticism, playing musical instruments, etc
What are NMDA receptors in interpositus necessary for? For what are they not?
- Necessary for *acquisition* (training) but NOT expression of eyeblink CR Pass the *necessity test* for *learning* but NOT for expression
Extrapyramidal Motor Pathway
- Not all DESCENDING motor signals originate in motor cortex -> a lot originate in *midbrain, brainstem, cerebellum* - Extrapyramidal motor system consists of descending pathways from: 1. Cerebellum = projects down spinal cord via *Rubrospinal, Olivospinal tracts* 2. Brainstem Vestibular nuclei = *Vestibulospinal tract* -> These pathways help to guide + modulate commands that originate in the *pyramidal* motor system 3. Brainstem Motor nuclei = *Reticulospinal tract* -> relay *"low level" innate + reflexive* motor commands tracts = bundle of axons that carry info to muscles from cerebellum and vestibular centers about balance + involuntary motor coordination
Development in Cerebellum
- Video of son who falls down walking over speed bump - Attempts to walk over it again and this time doesn't fall = Extrapyramidal motor pathways not fully developed in children, which control balance - practice + mistakes + attempts to learn build, modify, and strengthen these pathways
Where else do mossy fibers and climbing fibers project to?
1. *Mossy fibers* project from pontine to interpositus AND project to *granule cells* in cerebellar cortex -> granule projects onto *Purkinje* cells -> project back down to *inhibit* interpositus 2. *Climbing fibers* project from IO to interpositus AND project to *Purkinje cells* in cerebellar cortex!
Two types of Glutamate receptors:
1. AMPA 2. *NMDA* - BOTH a *VG* and *Ligand-gated* channel -- 1. Ligand gated = glutamate must bind to it -- 2. VG = Mg2+ ions like to sit in its channel & block it -> depolarization repels pos charged Mg and it pops out of channel
Long-term potentiation (LTP)
1. Ca entry through NMDARs activates molecule *calmodulin* (CAM) -> activates molecule *CAM kinase II* (CAMKII) 2. CAMKII triggers *chemical signaling cascade* = leads to *insertion of more AMPA receptors* into postsynaptic mem of *mossy fiber* synapse 3. Additional AMPARs make synapse *stronger* = *long-term potentiation* of synapse
Necessity test of NMDA receptors
1. If *AP5* onto interpositus while mammals given CS-US pairings -> animal *fails to acquire* eye blink CR 2. If training continues *without AP5* -> animal acquires normal eyeblink CR 3. After animal fully trained -> *AP5* onto interpositus does *NOT* interfere w performance of learned response - After CR learned, impaired NMDARs does not impair CR expression CONCLUSION: NMDA receptors are *required for acquisition of CR, but not expression* - It is the thing that builds the bridge! But once bridge is built, it's not needed anymore!
Unconditioned Eye-blink Reflex Pathway:
1. Mildly aversive *Unconditioned Stimulus* (US) - e.g. *Air puff* or *weak shock* delivered to skin of eyelid -> activates 1st order *somatosensory* receptor neurons of *ophthalmic* nerve* 2. Somato. neurons activated by US -> project to 2nd order in *trigeminal nucleus* = one of the cranial sensory nuclei -> Trigeminal neurons send *efferents* to 3rd order *motor* in *cranial facial motor* nucleus 3. Cranial facial motor nuc = contains motoneurons that send axons directly to *NMJ* of eyelid muscles -> produce *unconditioned response* (UR) of *blinking eye* US -> Ophthalmic nerve's somatosensory neuron -> Trigeminal nucleus -> cranial facial motor nuclei -> eyelid muscles blink
Four major stages of Pavlovian Conditioning
1. Pretraining (Baseline) 2. Training (Acquisition) 3. Testing 4. Extinction
What activates motor neurons?
1. Spinal interneurons + sensory neurons in DRG = provide motoneuron inputs mediating simple motor reflexes + central pattern generation 2. Pyramidal motor sys = provides motoneuron inputs that initiate + control voluntary movements 3. *Extrapyramidal motor sys* = modulate motoneurons by fine tuning movements - Cerebellum exerts a lot of influence
Post-training (testing phase) neural response to CS (tone)
1. When CS presented alone *after* training -> pontine nucleus fire -> release glutamate onto interpositus 2. Glutamate binds to both AMPA + NMDA receptors (now more AMPA than before) -> mossy fiber synapse STRONG enough to generate large EPSP that does *two* things: --1. Triggers AP in PS neuron -> generates eyeblink --2. Opens NMDA receptors at mossy fiber synapse (depolarization from #1)
When pontine is activated by CS tone:
1. via mossy fibers = triggers a *sequence* of granule cells to fire in cerebellum - one set active right when tone comes on, another a little later, and one particular set when puff occurs
Cerebellum is...
A *motor learning engine*! - Cerebellar circuitry regulates eyeblink conditioning, whereby *well-timed* eyeblink response (CR) is generated after a tone (CS)
Pavlovian CR is typically what kind of behavior?
A *reflex behavior* that animals INNATELY perform when they expect the US (they don't have to be taught behavior, *only the expectation*) -> Animal already knows how to blink eye when air is puffed (that's what it naturally wants to do), but conditioning helps it learn to do it before US due to its understanding of the CS+US relationship - reflex = behavior in which a single stim elicits a single response -> Conditioned response still fits this defn (CR = a reflex) Rabbit = vestibulo-ocular reflex
Pre- versus Post- training AP5
AP5 onto interpositus before training = blocks NMDARs -> prevents Ca entry into cell -> mossy fiber synapse cannot be strengthened by addition of new AMPAR AP5 after training -> synapses has already been strengthened by addition of new AMPARs -> no work left for NMDAR to do -> drug does NOT affect expression of eyeblink CR
Synaptic plasticity in interpositus
After conditioning: fact that *no* change in neural response to CS occurs BEFORE interpositus, and *change* occurs AFTER it = interpositus is the point at which change occurs!!! Scientifically: *Plastic synapse* at interpositus' *mossy fiber input* from pontine nuc - grows stronger as CS is repeatedly paired w US
Pre-training (Baseline) neural response to *US*
At *climbing fiber* synapse between *inferior olive + interpositus*: 1. US presented alone *before* training = Inferior Olive fire and *release glutamate* onto interpositus 2. Glutamate binds AMPA receptors = generates EPSP *large enough* to trigger *AP in interpositus* = b/c there are *lots* of AMPA receptors in climbing fiber synapse 3. Depolarization of interpositus = pops Mg2+ cork off NMDA receptors at *mossy fiber* synapse - *BUT receptor does NOT open b/c it is not binding glutamate*!!!
Pre-training (Baseline) neural responses to *CS*
At *mossy fiber synapse* between *pontine + interpositus*: 1. CS presented alone *before* training = *pontine* neurons fire and *release glutatmate* onto interpositus -> glutamate binds AMPA + NMDA, but only *AMPA receptors* pass current (NMDA receptors blocked by Mg2+) 2. *Sodium* entry through AMPA receptors generates *EPSP* = but *too small* to trigger AP in interpositus = b/c there are *very few* AMPA receptors in mossy fiber synapse = WEAK synapse
Training (Acquisition)
CS and US are *paired repeatedly* so subject *learns* that the CS *predicts* US Brain changes in some way so behavior is diff in future = *CS* will elicit a *conditioned response* The tone predicts the puff, so rabbit closes eye when hears tone before the puff
Testing
CS is presented *without* US and subject generates a CR Tone is presented without puff but rabbit still blinks
Extinction
CS is repeatedly presented *by itself* so that subject learns the *CS no longer predicts US* -> CR diminishes and finally disappears
Pavlovian Eyeblink Conditioning Graphs
CS: tone = starts before US and ends at same time as US US: shock/air = starts later and is very brief; ends at same time as CS Eyelid position PRE: natural curve; starts when US starts and remains until US is done Eyelid position POST = CR: natural curve; starts shortly after CS and before US; sustains until US done
Pathway for conditioned stimulus pathway
Conditioned stimulus pathway (tone): 1. Cochlear nucleus excites *pontine* nucleus 2. Pontine excites *interpositus* - *Mossy fibers* project from pontine -> interpositus 3. Interpositus excites *red nucleus* 4. Red excites *cranial facial motor neurons* 5. cranial facial motor *blink the eye* *US puff pathway*: ophthalmic nerve somato. neuron -> trigeminal nuc -> cranial facial motor -> eyelid muscles
Excited granule cells then...
Excite Purkinje cells: - Similar to granule cells, a *sequence* of purkinje cells activated when tone turns on & particular set active when puff occurs
Cerebellum ("Little Brain") function + subdivisions
Exerts *adaptive gain control* over spinal reflexes Protrudes from base Consists of TWO major subdivisions: 1. *Cerebellar cortex* (sits under main cerebrocortical brain) 2. *Deep nuclei* in *brainstem* - communicates w cerebellar cortex + sends motor signals down spinal cord and to cranial facial motor nuclei Primary FUNCTIONS: 1. Adaptive Motor control 2. Balance 3. Possibly other cognitive fncs
Active purkinje cells then... When does the rabbit blink?
Inhibit Interpositus and prevents CR When tone occurs, rabbit waits for right moment to blink: - Does not blink right away when CS is on because other circuit inhibits interpositus and prevents blink until the right moment (for US to arrive)
How does synaptic plasticity/change occur at pontine-interpositus synapses?
Interpositus needs to know more than what is going with the CS (tone) - ALSO needs info about US (puff): *Inferior Olive* IO relays *US* info to interpositus 1. *Trigeminal nucleus* of eye blink reflex pathway sends axons -> *Inferior Olive* 2. Inferior Olive (IO) sends *climbing fibers* axons -> *interpositus* nucleus
Where does LTP occur?
It occurs at the mossy fiber synapse - Before LTP, the postsynaptic interpositus dendrite has few AMPARs - After TLP, more AMPARs present = more depolarization - Does not need NMDA receptors to allow Calcium to enter cell
What trains the plastic synapse?
It's "trained" by *climbing fiber inputs* that relay info about US from inferior olive to interpositus IO is necessary to deliver info about US
Behavior (CR) is observed as a means to___________
Measure *expectation*! Regard as indicator that animal has learned to expect the CS
AP5
NMDA antagonist drug = blocks NMDA receptors
Plastic synapse + receptors of post-synaptic
Pre-synaptic = Pontine mossy fiber axon terminals - Releases *glutamate* Post-synaptic = Interpositus neuron - Glutamate binds to ionotropic receptors; *AMPA + NMDA* NMDA = both
Pavlovian Eyeblink Conditioning
Rabbit eyelid (*nicitating* membrane) conditioning procedure - Developed by Dorie Gormzano (1950's) 1. *Conditioned Stimulus* = auditory tone 2. *Unconditioned Stimulus* = airpuff/shock delivered to eyelid -> Goal = Pair CS with US How: 1. CS turns on 1 or 2 sec before US 2. CS coterminates (ends at same time) with US - US is very brief 3. After many *pairings* -> rabbit learns to close eye to CS *BEFORE* US comes on = *conditioned response*
Correlation test: Recordings of INTERPOSITUS neurons McCormick & Thompson (1984)
Recorded neurons in interpositus during eyeblink conditioning with electrodes: 1. Neurons did NOT respond to CS (tone) *before* conditioning -> acquired *neural response to CS* in tandem w acquisition of *behavioral CR* Over 3 days of pairing CS+US -> Eye blink CR gets BIGGER + neural responses in interpositus to CS also get BIGGER - Interpositus neurons start responding to tone CORRELATION: increased conditioned eye blink response to CS -> increased neural activity of interpositus
Building a bridge from ear to eyelid
Separate known pathways: 1. Auditory signal (e.g. CS) activates hair cells in cochlea -> trigger AP in spiral ganglion -> activate neurons in cochlear nucleus 2. Eyeblink reflex pathway from ophthalmic nerve somatosensory receptors -> trigeminal nucleus -> cranial facial motor nuclei -> eyelid muscles Need to determine neural pathway from *auditory centers -> eyelid muscles that generate CR when CS occurs* - & How it turns on
In what neurons of the pathway were conditioned neural responses observed?
Similar conditioned neural responses observed *downstream* from interpositus = *red and facial motor* NOT *upstream* from interpositus = cochlear or pontine nucleus -> in these areas, neurons responded to the CS both *before and after* conditioning so the response was *purely auditory*
What role NMDA receptors play in Hebb's rule?
They can function as Hebbian coincidence detectors For Hebb's rule to occur, we need a coincidence detector to detect when cells fire together - if firing together causes wiring together, we need to know when firing together will occur - NMDA detects this! -> needs two things: *coincidence of glutamate release AND post synaptic depolarization in order to open* - Pontine activation = glutamate; Interpositus activation = depolarization
Conditioned Response
When the rabbit learns to close eye to the CS, before US comes on - CR = a *prediction* of the unconditioned stimulus
Hebb's Rule
When two cells (connected to e/o) are active at *same time*, synapse bt them becomes *stronger* and stays stronger for *long time afterwards* ("cells that fire together wire together") - Strong connection = In future, if one becomes active, it's likely to activate other (reproduces prior activity pattern) Rule requires a biological mechanism for *coincidence detection* to signal when pre + postsynaptic cell are activated at same time - aka biological mech to produce hebbian plasticity
Brainstem Motor nuclei
Where *cranial motor* neuron *cell bodies* reside Motor axons exit brainstem through the *cranial nerves* to contact muscles in head + face Specifically, eyeblink reflex corresponds to the cranial facial nucleus in the brainstem (controls muscles of the face)