Motor Cognition

1
Motor Cognition

• Overview
• Many people believe processes used to plan and
  enact a movement can be used in problem solving
  and reasoning
• Moreover, the processes involved in perceiving
  action are also involved in movement
• This lecture will introduce key ideas involved in
  motor cognition and memory for action needed
• in planning, and producing our own actions
  anticipating, perceiving, and interpreting the
  actions of others
• In remembering actions

2
Motor Cognition

• Terminology
• Movement is a voluntary displacement of a body
  part in space
• Action a series of movements performed to
  achieve a goal

3
Motor Cognition

• Perception-action cycle
• Refers to the transformation of perceived
  patterns (often visually) into coordinated
  patterns of movement
• Examples. Returning a tennis ball, picking up a
  mug of coffee, walking on uneven terrain
• According to this perspective much of human
  behavior involves the interconnection of
  perception and movement
• The mediating link between perception and action
  is that a shared representation that is, the
  coding of perception and action is shared in the
  brain

4
Motor Cognition

• Motor processing in the brain
• Neuroanatomy
• Area M1, the primary cortex. Neurons in this
  region control fine motor movements, and send
  fibers out to muscles themselves
• Premotor area (PM) sets up program for motor
  sequences, and send input into M1
• Supplementary motor area (SMA) sets up and
  executes motor plans

5
Motor Cognition

• Motor processing in the brain
• Neuroanatomy
• Think of these 3 areas as a hierarchy. M1 is at
  the bottom of the hierarchy, and it enables
  specific movements PM higher up and it enables
  sets of movements at top SMA represents
  overarching plans of action

6
Primary motor cortex (M1)
Posterior parietal cortex
Supplementary motor cortex (SMA)
Premotor cortex
7
Motor Cognition

• Some evidence for roles of three areas
• Mushiake (1991)
• Recorded single-cell activity in monkeys in M1,
  PM, and SMA
• 2 conditions of interest were IT (internally
  triggered) and VT (visually triggered)
• In both conditions monkeys were required to touch
  3 pads on a panel in the IT condition monkeys
  needed to remember sequence and then touch the
  panels in the remembered order in the VT
  condition monkeys touched panels as they were
  illuminated

8
Motor Cognition

• Some evidence for roles of three areas
• M1 cells were active in the premovement and
  movement periods consistent with the hypothesis
  that M1 is involved in movement
• More SMA cells were active in the IT than the VT
  condition, consistent with the idea that SMA is
  important during planning since planning is more
  important in IT than VT condition
• More neurons were active in the VT than the IT
  condition during the premovement and movement
  periods

9
Motor Cognition

• Some evidence for roles of three areas
• Conclusions
• Movement planning and production involves a
  number of neuropsychological processes with
  different functions. These processes occur in
  different brain regions, and often occur
  simultaneously
• Planning and then producing a response involves
  different neural processes than responding to
  environmental cues

10
Motor Cognition

• Summary
• We have reviewed the role of just 3 brain regions
  involved in motor cognition
• Other regions are also involved with their
  involvement depending on the precise nature of
  the task in fact its been said that it takes
  the whole brain to make a cup of coffee

11
Motor Cognition

• Shared representation
• A considerable amount of data suggests that we
  can efficiently represent the actions made by
  other people (e.g., through viewing)
• It has been hypothesized this occurs (and
  brain-activation studies support this) because
  the representation of perceived action and
  produced action is shared
• These representations enable us to interpret the
  actions of others, respond appropriately, and
  efficiently learn how to physically produce
  actions that were viewed

12
Motor Cognition

• Mirror Neurons
• Mirror neuronsrefers to neurons that fire when
  organism (monkey) performs a specific grasping
  movement, and when that same grasping movement is
  observed being performed by experimenter or monkey

13
Motor Cognition

• Mirror Neurons
• Human evidence
• A variety of techniques have been used to provide
  evidence for motor neurons in humans
• A transcranial magnetic stimulation study showed
  increased excitability in the motor system during
  observation of actions performed by another
  person, but only for the brain regions involved
  in the muscles used by the other person

14
Motor Cognition

• Apraxia
• Definition
• An impaired ability to generate skilled actions
  that cannot be attributed to basic sensory or
  motor disturbance
• Generally conceptualized to reflect a disruption
  of a distributed praxis network
• Research has often focused on transitive actions
  -- actions that involve manipulation of a tool or
  object
• Examples use of a hammer, spatula,
• Apraxia frequently observed after neurological
  damage
• praxis network was thought to be left lateralized

15
Motor Cognition
Diagram of praxis network
16
Component Model ApproachRoy and Square, 1994
Roy, 1996
Sensory/Perceptual System
Visual/Gestural Information
Auditory/Verbal Information
Visual Tool/Object Information
Pantomime
Conceptual System
Knowledge of Action
Knowledge of Tool/Object Function
Production System
Response Selection
Image Generation
Delayed Imitation
Working Memory
Concurrent Imitation
Response Organization/Control
17
Component Model ApproachRoy and Square, 1994
Roy, 1996
Sensory/Perceptual System
Visual/Gestural Information
Auditory/Verbal Information
Visual Tool/Object Information
Main Responsibilities Analyzing visual gestural
information Identifying key features of tools
and objects for use
18
Component Model ApproachRoy and Square, 1994
Roy, 1996
Sensory/Perceptual System
Visual/Gestural Information
Auditory/Verbal Information
Visual Tool/Object Information
Conceptual System
Knowledge of Action
Knowledge of Tool/Object Function
Main Responsibilities Understanding tools,
objects, and gestures different types of
conceptual knowledge are dissociable
19
Motor Cognition

• Conceptual knowledge
• It appears that different types of conceptual
  knowledge associated with action are dissociable
  from each other as demonstrated in the next slides

20
Motor Cognition

• Conceptual knowledge
• Function knowledge vs manipulation knowledge
• Buxbaum and Saffran (2002) investigated function
  and manipulation knowledge in apraxic and
  non-apraxic patients with LHD (aside, px were
  apraxic to gestural tests including tests of
  pantomime)
• Function knowledge which two items are most
  similar in function (e.g., stapler, cellophane
  tape, pen)
• Manipulation knowledge which two items are most
  similar in manner of manipulation (e.g., piano,
  typewriter, stove)

21
Motor Cognition

• Conceptual knowledge
• Results apraxic patients were more impaired in
  manipulation test, but than function test
• Kellenbach et al. (2003) used PET to investigate
  brain activation associated with function and
  manipulation judgments
• Results showed that when participants made
  conceptual judgments about function, there was
  activation of a left network consisting of the
  ventral premotor cortex and the posterior middle
  temporal gyrus
• When participants made manipulation judgment an
  additional region, and additional region, the
  intraparietal sulcus, was activated

22
Motor Cognition

• Conceptual knowledge
• The intraparietal sulcus plays an important role
  in skilled object use (Heilman, 1993)

23
Motor Cognition

• Conceptual knowledge
• Visual-gestural knowledge
• Beauchamp (2002) in neuroimaging study showed
  that bilateral regions of the middle temporal
  cortex were activated when tool motion (i.e.,
  gestural motion) was viewed in comparison to
  human motion (i.e., person jogging on the spot)

24
Motor Cognition

• Conceptual knowledge
• Beauchamp (2003) in neuroimaging study showed
  that middle temporal gyrus responded more
  strongly to tool motion videos and point-light
  displays of tool motion

25
Motor Cognition

• Point-light displays (aside)
• Animals and humans move in ways that are
  distinctive and different from the way in which
  nonhumans move. These patterns of movement are
  called biological motion
• Johansson (1973) developed the point-light
  display technique to investigate movement.
• Small light sources attached wrists, knees,
  ankles, shoulders, and heads of actors who
  performed various movements (e.g., walking,
  running, dancing)

26
Motor Cognition

• Conceptual knowledge
• Park and Roy (in prep) showed that patients with
  LHD, but not RHD were strongly impaired on
  function and manipulation tests but that patients
  with LHD and RHD were impaired on tests of
  visual-gestural knowledge

27
Motor Cognition

• Conclusions
• These studies suggest that different types of
  conceptual knowledge associated with action are
  dissociable from each other
• Three types of knowledge have been studied in
  depth. These are knowledge of
• Function
• Manipulation
• Visual-gestural knowledge of tool motion

28
Motor Cognition

• Imitation in this model can be accomplished in
  two different ways
• 1. directly from perception to action and
• 2. indirectly through long-term memory
• Evidence to support this comes from studies which
  have shown
• The general observation that meaningless actions
  can be imitated accurately in cognitively
  unimpaired individuals
• Findings of dissociation between imitation and
  pantomime (e.g., Goldenberg Hagmann, 1997
  Ochipa et al., 1994) stronger lateralization to
  pantomime than to imitation
• (interpret on basis of model)

29
Motor Cognition

• What is acquired when we view purposeful action
• It would appear that we derive the goal of the
  action
• (e.g., see a person reaching hand across table,
  grasping a mug of coffee, and moving is toward
  lips would be described as drinking a cup of
  coffee. In other words viewed actions tend to be
  described in terms of the goal of the action

30
Motor Cognition

• Imitation
• Imitation -- ability to understand the intent of
  a viewed action and then to reproduce it
• This needs to be distinguished from mimicry,
  which is reproduction of a behavior without
  understanding (e.g., a parrot mimics human
  speech)
• Meltzoff and Moore (1977) showed that newborn
  infants can imitate viewed action (sticking out
  tongue opening mouth etc.)
• By 6 months of age infants can imitate actions on
  objects (e.g., shaking a rattle)

31
Motor Cognition

• Imitation
• As infants grow older deferred imitation
  abilities increase (i.e., memory for imitated
  action)
• data show that infants as young as 18 months
  appear to represent intentions of actions not
  just the action itself
• E.g., children who watched an actor try to pull
  apart a dumbbell but failed, were more likely to
  try and pull apart the dumbbell than if they
  watched a mechanical device attempt to pull apart
  a dumbbell

32
Motor Cognition

• Components of imitation
• Decety et al. (1997) in neuroimaging studies
  compared brain activation of subjects as they
  viewed meaningless actions either intending to
  recognize or to imitate the viewed action
• Additional brain regions activated when intending
  imitate meaningless actions supplementary motor
  area (SMA), the middle frontal gyrus, the
  premotor cortex, the anterior cingulate, and the
  superior and inferior parietal cortices in both
  hemispheres.

33
Motor Cognition

• Conclusion intentions (top-down) processes of
  participant influenced observation of action.
  Regions activated during observation also are the
  ones involved in action generation.
• Observing an action with the intention to perform
  that action involves regions similar to those
  used to generate the action
• when intending to recognize an action activated
  regions were the memory encoding structures (the
  parahippocampal gyrus)

34
Motor Cognition

• When actions are viewed separating intention
  from means a neuroimaging perspective
• Several people have proposed that actions are
  often understood in terms of the intentions
  (goals) they achieve and the means used
  (movements) to achieve these goals (e.g., Heider)
• Chaminade (2002) investigated using neuroimaging
  the neural regions associated with goals and
  means

35
Motor Cognition

• When actions are viewed separating intention
  from means a neuroimaging perspective
• Participants saw an actor make Lego
  constructions participants viewed the goal alone
  (hand moving away from block in specified
  location) the means alone (the hand grasping and
  moving the block) or the entire action. All
  participants imitated action just observed

36
Motor Cognition

• When actions are viewed separating intention
  from means a neuroimaging perspective
• Findings when participants imitated action or
  means, the medial prefrontal cortex was
  activated this region appears to play a critical
  role in inferring other peoples intentions
• When participants imitated goal the left premotor
  cortex activated
• Conclusion
• In normally functioning adults imitating a
  gesture activates neural regions associated with
  the intentions underlying the action

37
Motor Cognition

• Mental simulation
• Since imagery and perception activate similar
  brain regions it seems reasonable to hypothesize
  that one way to reason is to simulate (or
  imagine) the consequence of performing a planned
  action

38
Motor Cognition

• Simulation theories of action understanding
• It has been proposed that actions of others are
  understood by putting yourself in their place
  (either by observation or imagination)
• This permits you to derive their intentions and
  generate an action plan (but how can you do this
  since intentions and actions are internal and
  unobservable?)

39
Motor Cognition

• Mental simulation
• It has been shown that practicing with mental
  imagery can help participants in their
  performance of the task
• It has been shown that mental imagery practice
  has positive effects on complex motor skill
  learning (e.g., putting a golf ball)
• Yue Cole (1992) showed compared finger strength
  of two groups group 1 performed repeated
  isometric exercises group 2 received motor
  imagery instruction and imagined making finger
  movements without actually making them
• Both groups had increased finger strength group
  1 by 30 and group 2 by 22
• Conclusion possible to increase strength without
  actually repeated muscle activation

40
Motor Cognition

• Mental simulation
• It has also been shown that viewing an action can
  facilitate enactment of that action
• Priming refers to the facilitation of processing
  by previous performance of a task
• Motor priming refers to the facilitative effect
  that watching a movement or action has on making
  a similar motor response. Motor priming has been
  observed in a variety of experimental situations

41
Motor Cognition

• Mental simulation
• fMRI studies have shown that the neural
  difference between motor imagery and motor
  performance is not a matter of what, but how
  much
• i.e., similar brain regions are activated, but
  the level of activation is significantly lower
  in one study imagery activation was 30 of that
  found in actual execution (Roth, 1996)

42
Motor Cognition

• Mental simulation
• Ruby Decety (2001) Nature Neuroscience
  investigated the question of agency
• Backgroundif viewing an action activates regions
  involved in performing an action, how do people
  distinguish actions they perform vs those they
  observe (i.e., attribution of action to self or
  another agent)
• Previous studies have shown that the first-person
  perspective (imagining oneself) is associated
  with activation of inferior parietal, premotor,
  and SMA on left side

43
Motor Cognition

• Mental simulation
• This study asked what areas are activated when we
  imagine not ourselves performing an action, but
  another person performing that action
• Method
• Participants were scanned while they simulated
  everyday actions (e.g., winding a watch) with
  right hand (all Ps were right handed)
• Ps instructed to imagine themselves perform the
  action or to imagine another person performing
  the action
• Perspectives initiated by presenting a photo or
  from a spoken sentence describing the action

44
Motor Cognition

• Conclusions
• First person perspective versus imaging another
  person acting was associated with activation of
  common neural resources
• Consistent with notion that a common code is used
  to perceive, imagine, and produce actions
• However, specific regions are activated when
  imagining oneself performing an action versus
  another person. These regions may be used to
  determine agency

45
Motor Cognition

• Mental simulation
• Results
• Both self perspective and other perspective
  activated common regions supplementary motor
  area (SMA), premotor cortex, precuneus (an area
  located in parietal lobe), and occipital-temporal
  lobe
• However, when compared to the first-person
  perspective, the third-person perspective
  selectively activated the frontopolar cortex, the
  precuneus, and the right inferior portion of the
  parietal lobe
• See figure in next slide

46
Ruby Decety (2001)
47
Motor Cognition Memory

• Memory for action
• Subject-performed task (SPT) paradigm requires
  participant(P) to perform actions according to
  verbal instructions given by experimenter (e.g.,
  roll the ball, fold the paper, lift the pen) at
  study
• At test Ps memory for these actions is tested
• Control condition P hears instructions but does
  not perform actions
• Resultmemory for enacted action phrases is
  superior to that for events encoded without
  enactment

Presentation relies on Nilsson (2000) In Craik
and Tulving Oxford Handbook of memory
48
Motor Cognition Memory

• Memory for action--theories
• Non-strategic encoding theory of Cohen
• This theory proposed that enacted actions are
  encoded nonstrategically unlike verbal and other
  types of events

49
Motor Cognition Memory

• Memory for action--theories
• Multimodal theory of Backman and Nilsson (1984,
  1985)
• Enactment during encoding automatically leads to
  multimodal processing, which produces a rich
  encoding of information (multimodal because there
  is auditory, visual, and haptic input) in SPT
  condition
• subsequently proposed that physical (perceptual)
  properties were encoded nonstrategically, whereas
  verbal components were encoded strategically.
• SPTs contained verbal and physical properties
  whereas VTs contained verbal component only
  (Backman et al. 1986)

50
Motor Cognition Memory

• Memory for action--theories
• (Backman et al. 1989) proposed that the physical
  component of the dual code is encoded
  incidentally and retrieved implicitly, whereas
  verbal component is encoded intentionally and
  retrieved explicitly

51
Motor Cognition Memory

• Memory for action--theories
• Engelkamp and Zimmer (1984, 1985 etc.) proposed
  that encoding SPTs is governed by separate motor,
  visual, and verbal programs that produce separate
  modality-specific representations
• Motor encoding is more efficient than the other
  types of encoding and this results in the
  enactment effect

52
Motor Cognition Memory

• Memory for action--theories
• Motor coding improves item-specific encoding,
  whereas visual and verbal processing result in
  relational encoding between the items
• Two types of relational encoding 1. integration
  of actions within a list 2. integration of noun
  and verb in a command
• Hypothesized that type 1 integration is
  independent of enactment, and type 2 integration
  is hindered by enactment

53
Motor Cognition Memory

• Memory for actiondata in support of Cohen
• 1. no levels of processing effect
• 2. no primacy effect
• 3. no generation effect
• 4. no rate of processing effect
• 5. no difference in memory performance for
  children of different ages, for mentally retarded
  and controls or for elderly.
• These results all support notion that enactment
  (in contrast to verbal encoding) does not require
  strategic processing at encoding as hypothesized
  by Cohen

54
Motor Cognition Memory

• Memory for actiondata in support of multimodal
  coding theory
• These data are supported by studies that have
  used a divided attention as Ps encode SPTs at
  study (e.g., bounce the ball). At test memory for
  perceptual (color of object) and conceptual
  aspects (recall SPT) was tested (from Backman,
  Nilsson, Herlitz, Nyberg, Stigsdotter, 1991)
• Results shown in next slide

55
Motor Cognition Memory
56
Motor Cognition Memory

• Memory for actiondata in support of multimodal
  coding theory
• This next experiment very similar to previous
  one
• Ps encode SPTs at study (e.g., bounce the ball)
  under full and divided attention.
• At test memory for perceptual (weight of object)
  and conceptual aspects (recall SPT) was tested
  (from Backman, Nilsson, Herlitz, Nyberg,
  Stigsdotter, 1991)
• Results shown in next slide

57
Motor Cognition Memory
58
Motor Cognition Memory

• Conclusions
• recall of perceptual components is less strongly
  affected by DA than recall of the verbal
  instruction
• Supports hypothesis that verbal component
  requires strategic processing and physical
  (perceptual) component is more automatic or
  non-strategic
• Potential problem with experiment is that color
  recall is a form of cued recall, whereas verbal
  recall is a form of free recall. (This problem
  was addressed in Backman et. al. 1993).

59
Whats a tool?
a manipulable object that is used to transform
an actors motor output into a predictable
mechanical action for the purpose of attaining a
specific goal (Frey, 2007)

• Simple vs. complex tools

• Simple tools amplify the movement of the upper
  limbs (e.g., using a stick to extend reach)
• Complex tools provide a mechanical advantage and
  convert hand movements into qualitatively
  different actions (e.g., using scissors to cut
  paper)

60
What do I need to know to use this tool?
colour of the tool
function of the tool
manner of grasping the tool
identity of the recipient
how the tool physically is used
colour of the recipient
learned motor skill
61
Memory Systems
Relies on frontal striatal network
Gradual acquisition of skills
Implicit retrieval
Resistant to interference and decay
Relies on medial temporal structures
Semantic (i.e., facts) and episodic (i.e., recollection of events) memory
Conscious retrieval
Sensitive to interference and decay
knowing how
knowing what
62
Memory for Tools
63
Overview of Roy Park (2010)

• Investigated memory systems involved in the
  acquisition of different types of complex tool
  knowledge in a single study

• Examined extent to which an amnesic individual
  could acquire knowledge and skills related to
  novel complex tools

Why study amnesia?
- Individuals with amnesia are impaired in
acquiring new declarative knowledge, but have
intact procedural learning
Ideal population to study dissociation between
declarative and procedural aspects of tool
knowledge!
64
Method
Participants

• D.A.
• - 58 year old man with 17 years of
  education
• - Diagnosed with retrograde and
  anterograde amnesia after
• contracting herpes encephalitis in 1993
  

Neuroanatomical Profile Cognitive Functioning
Damaged / Impaired Medial temporal lobe structures (bilaterally), right anterior temporal lobe Delayed memory
Spared / Unimpaired Dorsal frontal, superior and inferior parietal, and posterior cingulate regions Immediate memory, visual naming, fluency, digit span, and executive functioning

• 6 healthy age and education-matched controls (3
  males, 3 females)

65
Method
Materials

• 15 novel unimanual complex tools constructed
  using
• KNEX

• Tools were designed to act on a recipient (e.g.,
  plastic wheel) to perform a specific function
  (e.g., move wheel down a path)

• Tool function, manner of grasping, or manner of
  use cannot be inferred based on physical
  appearance

66
Example of Novel Complex Tool
67
Procedure

• Each session (S1, S2, S3) had 3 phases
• 1) Pre-test
• 2) Training
• 3) Post-test

68
Procedure
1) Pre-test - Recall test (e.g. tool
function, tool colour) - Recognition
test - Grasp-to-command -
Use-to-command
2) Training Phase - 2 blocks (10 target
tools x 2) - Video demonstration
followed by practice - Limit of 90
seconds to complete one errorless trial
- Experimenter provided feedback
3) Post-test (same format as Pre-test)
69
Procedure

• Task order remained the same across sessions
  except....

D.A.s S3 Post-test

• Recall test
• Recognition test

• Grasp-to-command
• Use-to-command
• Use-to-command Recipient cued (RC) trial

• Grasp-to-command
• Use-to-command

• Use-to-command Recipient cued (RC) trial

Changes made to bring D. A.s performance off
the floor
70
Hypotheses
1) D.A. would demonstrate unimpaired motor skill
acquisition associated with novel complex
tools (i.e., becoming faster in using the
tools across training trials)
2) D.A. would be impaired in his ability to
recall the properties (functional and
perceptual) of the novel tools
3) D.A. would be impaired on tasks that required
him to consciously demonstrate the
appropriate grasp and trained use of the
novel complex tools.
4) There would be no effect of the 3-week delay
on measures of procedural memory in either
D.A. and controls, but that there would be
an effect of the 3-week delay on measures
of declarative memory in the controls
71
Training

• No differences between D.A. and controls in any
  training trial

• Completion time decreased by approximately 3.4
  seconds per trial in controls and 6.3 seconds per
  trial in D.A.

• No effect of the 3-week delay found in either
  D.A. or controls

72
Recall
(3 days) (3 weeks)
(3 days) (3 weeks)

• For functional associative recall, D.A.s
  performance was worse than controls in all trials
  except in S3 Post

• For perceptual recall, D.A.s performance was
  worse than controls in all trials except in S3
  Pre and S3 Post

• Performance for controls is significantly worse
  after the 3-week delay for both categories

73
Grasp-to-command
(3 days) (3 weeks)

• D.A.s grasp-to-command accuracy was worse than
  controls only in S3 Post

• Grasp-to-command accuracy in controls was worse
  after the 3-week delay

74
Use-to-command

• D.A.s completion time was worse than controls
  at S2 Post and S3 Pre, and his accuracy was worse
  than controls at all trials

• D.A.s performance on both measures in the RC
  trial was better with the target tools than the
  lure tools

• Controls were slower and less accurate after the
  3-week delay

75
Conclusion