COGS 107B - Winter 2010 - Lecture 15 - Memory


Published on

COGS 107B - Winter 2010 - Lecture 15 - Memory

  • Be the first to comment

  • Be the first to like this

No Downloads
Total views
On SlideShare
From Embeds
Number of Embeds
Embeds 0
No embeds

No notes for slide

COGS 107B - Winter 2010 - Lecture 15 - Memory

  1. 1. Cogs 107b – Systems Neuroscience lec15_03022010 learning and memory II principle: ‘implicit vs. explicit memory’ the term ‘procedural knowledge’, sometimes called implicit memory, denotes knowledge of how to accomplish a task, and often pertains to knowledge which unlike ‘declarative knowledge’ cannot be easily articulated by the individual, or knowledge that is nonconscious.
  2. 2. spike-timing (pre-post) synaptic change spike-timing (pre-post) spike-timing (pre-post) spike-timing (pre-post) synaptic change synaptic change synaptic change no ACh, no NE (as in NREM sleep) ACh + NE (as in waking) NE, no ACh (??) ACh, no NE (e.g., REM sleep) co-activity rules for synaptic efficacy change depend on neuromodulatory systems spike-timing (post-pre) spike-timing (post-pre) spike-timing (post-pre) spike-timing (post-pre)
  3. 3. learning 101: Pavlovian fear conditioning – association of an initially neutral tone (CS) with a foot-shock results in ‘freezing’ responses when the tone alone is played – this type of memory may be implicit or explicit lateral amygdala central amygdala auditory cortex / auditory thalamus (neurons excited by sound stimulus - CS) somatosensory cortex / somatosensory thalamus (neurons excited by shock stimulus - US) before pairing after pairing freezing response, increased heart rate, etc. hippocampus (place information) IT (visual item information) conditioned stimulus (CS) conditioned + unconditioned stimulus (US) conditioned stimulus curiosity ‘ shock and awe’ (freezing) conditioned freezing
  4. 4. reach onset pellet contact muscle 1 muscle 2 muscle 3 muscle 4 muscle 5 muscle 6 muscle 7 Kargo and Nitz, JNS, 2003 mod. from Conner et al., Neuron, 2003 early lrng. late lrng. w/out ACH w/out ACH w/out ACH rats learn, across days, to efficiently reach and grasp a small sugar pellet over those days, the muscle patterns used in grasping adapt over those days, the area of primary motor cortex taken up by neurons associated with the reaching limb grows if ACh inputs to the primary motor cortex are removed, neither the learning nor the changes in motor cortex occur implicit/procedural learning I: rat reach-to-grasp as a model of motor skill learning - ‘more is better’ and ACh helps
  5. 5. implicit/procedural learning II: perceptual skill learning - ‘more is better’ – ACh helps control trained rats trained to make a nosepoke if they detect a 4 kHz tone show improvements in detection over days of training Over the same time period the topographic representation of pitch in primary auditory cortex changes such that more neurons respond to 4 kHz tones in separate experiments, pairing of a 9 kHz tone with stimulation of ACh neurons in the basal forebrain changes the topographic representation in primary auditory cortex such that more neurons respond to 9 kHz tones Polley et al., JNS, 2006 Kilgard et al., Science, 1998
  6. 6. implicit/procedural learning III: basal ganglia versus hippocampus – an ACh-based competition in training (below), the rat is taught to move to the goal to obtain reward subsequently, on test trials (above), the maze is turned upside-down and the rat demonstrates whether he has learned to ‘make a left’ at the ‘T’ (a response strategy) or to ‘move to that place in the room’ (a place strategy) if the rat is asked this question early in training (within the first couple of days), one tends to see a ‘place’ strategy and ACh is high in the hippocampus if the rat is asked this question late in training, one tends to see a ‘response’ strategy and ACh is high in the basal ganglia early in training, when one would normally expect a ‘place’ strategy, inactivation of the hippocampus (the home of ‘place cells’) results in the emergence of a response strategy late in training, when one would normally expect a ‘response’ strategy, inactivation of the basal ganglia (proposed to select responses via the direct pathway) results in the emergence of a place strategy thus, the animal has learned two separate strategies which compete for expression
  7. 7. episodic memory (memory for events and their ordering - a form of explicit memory): hippocampal cell activity in a ‘place’ often depends on the places previously or subsequently visited (this is termed retrospective and prospective place coding) for one block of trials, the animal must travel to the west end when placed at either the N or S start point for the next block, the animal must travel to the east end when placed at either the N or S start point about half of all hippocampal neurons fire spikes (green dots) when the animal is in a certain part of the maze (here the S arm) – this is seen irrespective of the direction taken after reaching the middle other hippocampal neurons are ‘prospective’ – they fire in a certain place depending on where the animal will go from that place – they too have activity dependent on the full episode some hippocampal neurons fire in a certain place, but only if they reached that place from the N as opposed to the S side – their activity depends on the character of the full episode and is termed ‘retrospective’ adap. from Ferbinteanu and Shapiro, Neuron, 2003
  8. 8. adap. from Chafee and Goldman-Rakic, JNP, 1998 working memory: holding items in memory (7+/-2) is achieved through interaction of the prefrontal and parietal cortex parietal cortex prefrontal cortex delay period cue period saccade parietal neurons prefrontal neurons cue +delay delay +action
  9. 9. an example: prefrontal ‘top-down’ influences on parietal cortex during an oculomotor delayed response task – inactivation of prefrontal cortex via cooling depresses ‘working memory’ responses of parietal cortex neurons and increases errors error rate (percent saccades to wrong site) for each of 8 directions used increases when prefrontal cortex is inactivated parietal neuron has delay-period activity specific to the N and NW targets delay-period activity for the same neuron is depressed when prefrontal cortex is inactivated Chafee and Goldman-Rakic, JNP, 2000