Interactions Between Forms of Memory: When Priming Hinders New Episodic Learning

1
Interactions Between Forms of MemoryWhen
Priming Hinders New Episodic Learning

• Anthony D. Wagner, Anat Maril, and Daniel L.
  Schacter
• Journal of Cognitive Neuroscience, 2000

Wei-Chung Lee April 25, 2001
2
Forms of Memory
Long Term Memory
Explicit (Declarative)
Implicit (Non-declarative)
Semantic (Facts)
Episodic (Events)
Priming
Procedural
Associative Learning
Nonassociative Learning
Skeletal musculature
Emotional responses
Medial temporal lobe
Reflex pathways
Amygdala
Striatum
Neocortex
Cerebellum
3
Motivation

• When and how do these different forms of memory
  interact?

?
4
Episodic Memory

• Explicit memory for conscious recollection of
  events from past experience

Buckner and Koustaal, 1998
5
Perceptual Repetition Priming

• Facilitation derived from repeated exposure to a
  stimulus

Buckner et al., 1998
6
Hypothesis

• Priming for past experiences can hinder new
  episodic encoding.

7
Predictions

• Enhanced behavioral and neural priming during
  re-encoding with a shorter (2 min.) relative to a
  longer (25 hr.) lag between two encoding
  episodes.
• Impaired long-term retention of episodic memory
  in conditions with greater behavioral and neural
  priming.
• Lag Effect shorter lags between repetitions
  deleterious to explicit memory (Ebbinghaus, 1885).

8
Experimental Procedure

• Varied temporal lag between initial and repeated
  encoding episodes with words.

9
Behavioral Priming

• Mean median Reaction Times (RTs) for
  Abstract/Concrete decision (msec).
• Once-presented 777
• Long-Lag 763
• Short-Lag 720

p lt 0.09 p lt 0.0001
10
Neural Correlates of Episodic encoding and
Repetition Priming
Once-presented gt Fixation
Once-presented gt All Re-encoding

• A subset of regions activated by episodic
  encoding demonstrate a significant repetition
  priming effect (i.e.. decrease in activation)
  anterior-LIPC, posterior-LIPC, and left fusiform
  cortex.

11
Lag Effects on ROI Activation

• Long-Lag gt Short-Lag
• Priming also effects activation in left frontal
  operculum, left middle frontal, and medial
  frontal regions (Data not shown)
• Greater neural correlate of priming during
  re-encoding following a shorter temporal lag
  rather than long

12
Explicit Memory

• Explicit memory scores (pHit - pFalse alarm)
• Overall (High Low Confidence) pHit
• Long-Lag 35 77
• Short-Lag 31 73
• Once-Presented 22 64
• High Confidence explicit memory
• Long-Lag 32 46
• Short-Lag 25 38
• Once-Presented 17 30

p lt 0.005 p lt 0.0001
13
Priming Negatively Correlates with Subsequent
Explicit Memory

• A negative association across-subjects between
  behavioral neural (posterior anterior-LIPC
  shown) priming and subsequent explicit memory.
• Reliable correlations not found for Short-Lag
  trials (highest levels of priming led to lowest
  levels of High Confidence gt range too small)

14
Conclusions

• Correlation of behavioral and neural measures of
  priming during the re-encoding of a stimulus and
  impaired episodic encoding.
• cross-talk between implicit and explicit forms
  of memory.
• Posited mechanism Priming reduces encoding
  variability or multiple retrieval routes
  increasing the probability that same
  task-relevant stimulus features ...

15
Implications Questions

• Consolidation? Time course of Lag Effect
  optimal time to re-encode?
• Attention?
• 1 ?L neocortex ? 1x106 neurons ? 1x109 synapses
  and 1 voxel ?
• Segregation of memory types justified?
• Activation maps show significant differences in
  Blood Oxygenation Levels.
• Encoding versus re-encoding
• Not encoding/re-encoding the same words between
  sessions - may evoke different meanings/salienci
  es.