University of Dundee Metacognition Lectures 2024 PDF

University of the Year for Student Experience (The Times/Sunday Times Good University Guide 2020) Metacognition Dr Chris Benwell 2024 dundee.ac.uk...

University of the Year for Student Experience (The Times/Sunday Times Good University Guide 2020) Metacognition Dr Chris Benwell 2024 dundee.ac.uk Page 1 Metacognition - Lecture Series 1. Introduction to metacognition 2. Measuring metacognition 3. Metacognition and psychopathology 4. Metacognition and belief structures 5. Neural correlates of metacognition 6. Evolution and metacognition in other species 7. Presentations/discussions 8. Metacognition and consciousness 9. Improving metacognition 10. The limits of self-knowledge dundee.ac.uk Page 2 Metacognition – Assessment Level 4 Coursework: 1 x 2500-word assignment due on Wednesday November 13th. Describe and critically evaluate two studies which suggest that relationships exist between metacognition and other psychological traits such as belief structures and/or dimensions of psychopathology. Level 4 Exam: 2 hour, on-campus exam at end of semester. Worth 60% of module grade. Level 5 Coursework: 2 x 2500-word assignments, each worth 50% of the module grade: 1st assignment due: Friday 15th November (12 noon) 2nd assignment due: Friday 29th November (12 noon) dundee.ac.uk Page 3 Metacognition – Assessment Level 5 Coursework: 2 x 2500-word assignments, each worth 50% of the module grade. 1st assignment due: Friday 15th November Describe and critically evaluate two studies which suggest that relationships exist between metacognition and other psychological traits such as belief structures and/or dimensions of psychopathology. 2nd assignment due: Friday 29th November Does metacognition depend on mechanisms and information that are separable from those underlying cognitive performance itself (i.e., is metacognition a 2 nd- order process)? Consider evidence from behavioural, neuroimaging and/or lesion studies dundee.ac.uk Page 4 What to study? (1) Lecture notes (2) I will upload relevant papers for each lecture to the module page on MyDundee. (3) Any questions/concerns: [email protected] dundee.ac.uk Page 5 Consciousness State of being aware of our surroundings and internal experiences. dundee.ac.uk Page 6 Consciousness State of being aware of our surroundings and internal experiences. Usually associated with states of high arousal and wakefulness dundee.ac.uk Page 7 Consciousness State of being aware of our surroundings and internal experiences. Usually associated with states of high arousal and wakefulness An experience or other mental entity is 'phenomenally conscious' just in case there is 'something it is like' for one to have it. dundee.ac.uk Page 8 The hard problem A fundamental mystery in neuroscience and psychology is subjective, 1st-person conscious experience (qualia). dundee.ac.uk Page 9 The hard problem A fundamental mystery in neuroscience and psychology is subjective, 1st person conscious experience (qualia). No intelligible causal relationship between physiological states and qualia has ever been proposed. dundee.ac.uk Page 10 The hard problem A fundamental mystery in neuroscience and psychology is subjective, 1st person conscious experience (qualia). No intelligible causal relationship between physiological states and qualia has ever been proposed. Even if I knew every physical action underlying your experience of pain or of seeing the colour red, it seems I would still miss entirely your subjective sensations associated with the experience. dundee.ac.uk Page 11 The hard problem A fundamental mystery in neuroscience and psychology is subjective, 1st person conscious experience (qualia). No intelligible causal relationship between physiological states and qualia has ever been proposed. Even if I knew every physical action underlying your experience of pain or of seeing the colour red, it seems I would still miss entirely your subjective sensations associated with the experience. Qualia appear not to be objectively measurable. dundee.ac.uk Page 12 The hard problem A fundamental mystery in neuroscience and psychology is subjective, 1st person conscious experience (qualia). Thomas Nagel dundee.ac.uk Page 13 The hard problem A fundamental mystery in neuroscience and psychology is subjective, 1st person conscious experience (qualia). In "What Is It Like to Be a Bat?", Nagel writes that science cannot describe what it is like to be a thinker who conceives of the world from a particular subjective perspective. Thomas Nagel dundee.ac.uk Page 14 The hard problem A fundamental mystery in neuroscience and psychology is subjective, 1st person conscious experience (qualia). Functionalism: the identity of a mental state is determined by its causal relations to sensory stimulations, other mental states, and behavior. dundee.ac.uk Page 15 The hard problem A fundamental mystery in neuroscience and psychology is subjective, 1st person conscious experience (qualia). Functionalism: the identity of a mental state is determined by its causal relations to sensory stimulations, other mental states, and behavior. A functionalist might characterize pain as the state that tends to be caused by bodily injury, to produce the belief that something is wrong with the body and the desire to be out of that state, to produce anxiety, and, in the absence of any stronger, conflicting desires, to cause wincing or moaning. dundee.ac.uk Page 16 The hard problem A fundamental mystery in neuroscience and psychology is subjective, 1st person conscious experience (qualia). Critic of ‘functionalism’: the proposal that mental states should be explained with reference to their functional role (consciousness as software). Ned Block dundee.ac.uk Page 17 The hard problem A fundamental mystery in neuroscience and psychology is subjective, 1st person conscious experience (qualia). Critic of ‘functionalism’: the proposal that mental states should be explained with reference to their functional role (consciousness as software). A system with the same behavioural and Ned Block functional states as a human need not be conscious. dundee.ac.uk Page 18 The hard problem This fundamental disconnect between the objective substrates and subjective quality of conscious experience has been labelled the “hard problem” of consciousness (Chalmers, 1997). David Chalmers dundee.ac.uk Page 19 The hard problem A fundamental mystery in neuroscience and psychology is subjective, 1st person conscious experience (qualia). David Chalmers Thomas Nagel Ned Block Rene Descartes dundee.ac.uk Page 20 The hard problem A fundamental mystery in neuroscience and psychology is subjective, 1st person conscious experience (qualia). David Chalmers Thomas Nagel Ned Block Rene Descartes Mind-body problem, most famously defined by Descartes, has not yet been solved. dundee.ac.uk Page 21 Why is the hard problem so hard? The dominant scientific paradigm at the dawn of modern science (and during Descartes’ life) was the ‘Mechanical Philosophy’. dundee.ac.uk Page 22 Why is the hard problem so hard? The dominant scientific paradigm at the dawn of modern science (and during Descartes’ life) was the ‘Mechanical Philosophy’. The natural world is made up of matter in motion, and natural phenomena can be explained by the structure of these bodies and their action on one another. In other words, the universe is like a complicated machine and is best understood in this way. dundee.ac.uk Page 23 Why is the hard problem so hard? The dominant scientific paradigm at the dawn of modern science (and during Descartes’ life) was the ‘Mechanical Philosophy’. The natural world is made up of matter in motion, and natural phenomena can be explained by the structure of these bodies and their action on one another. In other words, the universe is like a complicated machine and is best understood in this way. dundee.ac.uk Page 24 Why is the hard problem so hard? The dominant scientific paradigm at the dawn of modern science (and during Descartes’ life) was the ‘Mechanical Philosophy’. The natural world is made up of matter in motion, and natural phenomena can be explained by the structure of these bodies and their action on one another. In other words, the universe is like a complicated machine and is best understood in this way. Rene Descartes Thomas Hobbes dundee.ac.uk Page 25 Newton: action at a distance Isaac Newton dundee.ac.uk Page 26 Newton: action at a distance Isaac Newton A contemporary of Descartes, Newton showed that the mechanical philosophy does not hold for all natural phenomena. Specifically, it could not account for the force of gravity. dundee.ac.uk Page 27 Newton: Action at a distance Isaac Newton A contemporary of Descartes, Newton showed that the mechanical philosophy does not hold for all natural phenomena. Specifically, it cannot account for the force of gravity. Potentially of relevance for the hard problem, Newton demonstrated that the mechanical philosophy's emphasis on contact as the only efficient cause was insufficient to explain the natural world. dundee.ac.uk Page 28 Action at a distance Since Newton, other natural phenomena have been discovered which cannot be explained within the Mechanical Philosophy including observations in quantum physics, chemistry and biology However, despite these established limitations, the mechanical philosophy continues to underpin many fields of modern science including neuroscience. Is this a historical artifact? Or is mechanical causation the only form of causation the human mind can comprehend? dundee.ac.uk Page 29 The hard problem Some theorists have suggested that the hard problem might not be solvable. Conscious experience might permanently lie outside of human understanding. Colin McGinn Noam Chomsky dundee.ac.uk Page 30 The hard problem and metacognition Discuss: Do you think the hard problem of consciousness can be solved? How might metacognition research contribute? dundee.ac.uk Page 31 The hard problem and metacognition Megan Peters (2022) has recently proposed that research into metacognition might offer a small step towards tackling the hard problem of conscious experience. dundee.ac.uk Page 32 The hard problem and metacognition Megan Peters (2022) has recently proposed that research into metacognition might offer a small step towards tackling the hard problem of conscious experience. She labelled this approach Metacognition as a Step Toward Explaining Phenomenology (M-STEP) dundee.ac.uk Page 33 The hard problem and metacognition With M-STEP, Peters proposes that there are certain properties of metacognition which might allow us to build a relatively complete model of the computations which give rise to the subjective experience of confidence. dundee.ac.uk Page 34 The hard problem and metacognition With M-STEP, Peters proposes that there are certain properties of metacognition which might allow us to build a relatively complete model of the computations which give rise to the subjective experience of confidence. dundee.ac.uk Page 35 The hard problem and metacognition David Marr (1982) famously argued that any information processing system can be analysed at three levels, that of → (1) the computational problem the system is solving; → (2) the algorithm the system uses to solve that problem; and → (3) how that algorithm is implemented in the “physical hardware” of the system (for us that is probably in the brain). dundee.ac.uk Page 36 The hard problem and metacognition dundee.ac.uk Page 37 The hard problem and metacognition But could a generative model of subjective confidence be generalised to other phenomenological experiences? “…it is the generative explanation of phenomenal character that is the target of our investigation, not the output of that generative process (i.e., not the specific qualia themselves or the fact that they possess phenomenal character).” Peters (2022) dundee.ac.uk Page 38 The hard problem and metacognition Strong form argument: All phenomenological/subjective experiences are generated by the same neurocomputational process: a canonical computation for phenomenal experiences dundee.ac.uk Page 39 The hard problem and metacognition Strong form argument: All phenomenological/subjective experiences are generated by the same neurocomputational process: a canonical computation for phenomenal experiences → By building a model of perceptual confidence (including perhaps its neurocomputational substrates), we could identify a first step in the generative process that gives rise to the phenomenal character of all subjective experiences. dundee.ac.uk Page 40 The hard problem and metacognition Intermediate form argument: Type-1 subjective experiences (C1/primary: red, pain, etc.) are generated by different neurocomputational processes than type-2 subjective experiences (C2/reflective: confidence, uncertainty, clarity). dundee.ac.uk Page 41 The hard problem and metacognition Intermediate form argument: Type-1 subjective experiences (C1/primary: red, pain, etc.) are generated by different neurocomputational processes than type-2 subjective experiences (C2/reflective: confidence, uncertainty, clarity). → By building a model of perceptual confidence (including perhaps its neurocomputational substrates), we will still identify a first step in the process that gives rise to one category of phenomenology: the reflective type (C2). dundee.ac.uk Page 42 The hard problem and metacognition Weak form argument: Every phenomenological/subjective experience is generated by its own unique neurocomputational process. dundee.ac.uk Page 43 The hard problem and metacognition Weak form argument: Every phenomenological/subjective experience is generated by its own unique neurocomputational process. → This would mean there is no such thing as a canonical computation common across any two subjective experiences, no common substrate for the type of experience that has phenomenal character because there is no single type of experience with the common property of ‘has qualia’. dundee.ac.uk Page 44  The hard problem and metacognition Potential criticisms of M-STEP: → The approach still leaves out exactly why certain qualia (i.e. redness, pain) are associated with certain subjective experiences (and not others). dundee.ac.uk Page 45  The hard problem and metacognition Potential criticisms of M-STEP: → The approach still leaves out exactly why certain qualia (i.e. redness, pain) are associated with certain subjective experiences (and not others). → Model Parameters might still not be meaningfully linked to the actual contents of subjective experience in any way. dundee.ac.uk Page 46  The hard problem and metacognition Potential criticisms of M-STEP: → The approach still leaves out exactly why certain qualia (i.e. redness, pain) are associated with certain subjective experiences (and not others). → Model Parameters might still not be meaningfully linked to the actual contents of subjective experience in any way. → Is it possible for us to discover all the necessary model inputs and outputs through typical experimental designs? Some might not be reportable by participants. dundee.ac.uk Page 47 The hard problem and metacognition Can confidence-based task measures be used to assess whether participants perceive stimuli consciously or unconsciously? Matthias Michel (2022) argues yes (but with caveats) dundee.ac.uk Page 48 The hard problem and metacognition Can confidence-based task measures be used to assess whether participants perceive stimuli consciously or unconsciously? Matthias Michel (2022) argues yes (but with caveats) → Based on the assumption that the metacognitive system only has access to sensory input that reaches conscious awareness but is blind to inputs that are processed by the sensory system but don’t reach conscious awareness. dundee.ac.uk Page 49 The hard problem and metacognition Can confidence-based task measures be used to assess whether participants perceive stimuli consciously or unconsciously? Matthias Michel (2022) argues yes (but with caveats) → Based on the assumption that the metacognitive system only has access to sensory input that reaches conscious awareness but is blind to inputs that are processed by the sensory system but don’t reach conscious awareness. → Under this assumption, meta-d’ (or the meta-d’/d’ ratio) can be thought of as indexing the level of conscious access the participant has to the sensory information driving their responses. dundee.ac.uk Page 50 The case of blindsight Blindsight is characterized by residual visual abilities in the absence of reported visual awareness following lesions to the primary visual cortex (Weiskrantz, 2009) dundee.ac.uk Page 51 The case of blindsight Blindsight is characterized by residual visual abilities in the absence of reported visual awareness following lesions to the primary visual cortex (Weiskrantz, 2009) dundee.ac.uk Page 52 The case of blindsight In post-decision wagering, the rate of advantageous bets falls dramatically for subthreshold stimuli in a blindsight patient, despite performing well above chance on a discrimination task (Persaud et al., 2007; Persaud et al., 2011). dundee.ac.uk Page 53 The case of blindsight In post-decision wagering, the rate of advantageous bets falls dramatically for subthreshold stimuli in a blindsight patient, despite performing well above chance on a discrimination task (Persaud et al., 2007; Persaud et al., 2011). dundee.ac.uk Page 54 The case of blindsight In post-decision wagering, the rate of advantageous bets falls dramatically for subthreshold stimuli in a blindsight patient, despite performing well above chance on a discrimination task (Persaud et al., 2007; Persaud et al., 2011). The difference between sighted and “blind” hemifields was a difference in metacognitive sensitivity, not visual sensitivity. dundee.ac.uk Page 55 Unconscious stimuli boost performance but not metacognition Similar effects have also been shown in healthy participants Vlassova et al., (2014) showed that unconscious stimulus information influenced first-order perceptual accuracy but not metacognitive sensitivity. dundee.ac.uk Page 56 Unconscious stimuli boost performance but not metacognition Similar effects have also been shown in healthy participants Vlassova et al., (2014) showed that unconscious stimulus information influenced first-order perceptual accuracy but not metacognitive sensitivity. dundee.ac.uk Page 57 The hard problem and metacognition Can confidence-based task measures be used to assess whether participants perceive stimuli consciously or unconsciously? Criticisms of Michel’s position dundee.ac.uk Page 58 The hard problem and metacognition Can confidence-based task measures be used to assess whether participants perceive stimuli consciously or unconsciously? Criticisms of Michel’s position → Factors other than conscious awareness are known to influence metacognitive efficiency. Hence it is difficult to attribute any lack of metacognitive efficiency to a lack of conscious awareness. dundee.ac.uk Page 59 The hard problem and metacognition Can confidence-based task measures be used to assess whether participants perceive stimuli consciously or unconsciously? Criticisms of Michel’s position → Factors other than conscious awareness are known to influence metacognitive efficiency. Hence it is difficult to attribute any lack of metacognitive efficiency to a lack of conscious awareness. → Experiments would need to be tightly and appropriately controlled to make inferences about consciousness using metacognitive efficiency. dundee.ac.uk Page 60 The hard problem and metacognition Can confidence-based task measures be used to assess whether participants perceive stimuli consciously or unconsciously? Criticisms of Michel’s position → Factors other than conscious awareness are known to influence metacognitive efficiency. Hence it is difficult to attribute any lack of metacognitive efficiency to a lack of conscious awareness. → Experiments would need to be tightly and appropriately controlled to make inferences about consciousness using metacognitive efficiency. → Metacognitive inefficiency may only indicate unavailability of information for metacognition, not unconsciousness dundee.ac.uk Page 61 The hard problem and metacognition Can confidence-based task measures be used to assess whether participants perceive stimuli consciously or unconsciously? Criticisms of Michel’s position → Factors other than conscious awareness are known to influence metacognitive efficiency. Hence it is difficult to attribute any lack of metacognitive efficiency to a lack of conscious awareness. → Experiments would need to be tightly and appropriately controlled to make inferences about consciousness using metacognitive efficiency. → Metacognitive inefficiency may only indicate unavailability of information for metacognition, not unconsciousness → The M-ratio (used to index metacognitive efficiency) has been shown to vary as a function of overall confidence bias (it is lower in states of low confidence) dundee.ac.uk Page 62 The hard problem and metacognition Summary: Michel’s position → Confidence based measures are a valid proxy for conscious awareness dundee.ac.uk Page 63 The hard problem and metacognition Summary: Michel’s position → Confidence based measures are a valid proxy for conscious awareness → However, we cannot be sure that they are accurate dundee.ac.uk Page 64 The hard problem and metacognition Summary: Michel’s position → Confidence based measures are a valid proxy for conscious awareness → However, we cannot be sure that they are accurate → Despite their limitations, Michel argues they are better than current alternatives for measuring conscious awareness. dundee.ac.uk Page 65 The hard problem and metacognition “Galileo's telescope-based procedure for detecting craters on the moon was extremely inaccurate compared to the procedures we would use today to achieve the same goal. But it was accurate enough. Galileo would have reached the same conclusion using a more accurate procedure. The inaccuracy of his procedure did not undermine his goals. Similarly, confidence-based procedures are most certainly inaccurate. But one can create conditions where they are accurate enough for our scientific goals. Again, whether this is the case or not for confidence-based procedures can only be decided on a case-by-case basis” dundee.ac.uk Page 66 The hard problem and metacognition Discuss: Can you think of times when metacognition dissociates from conscious awareness? Do you ever feel confident/not confident about something without awareness of why you feel that way (i.e., without an internal representation on which the confidence is based?)? dundee.ac.uk Page 67  Any questions? dundee.ac.uk Page 68 Next week Improving metacognition dundee.ac.uk Page 69 What to study? References Peters, M. A. (2022). Towards characterizing the canonical computations generating phenomenal experience. Neuroscience & Biobehavioral Reviews, 104903. Michel, M. (2022). Confidence in consciousness research. Wiley Interdisciplinary Reviews: Cognitive Science, e1628. Persaud, N., McLeod, P., & Cowey, A. (2007). Post-decision wagering objectively measures awareness. Nature Neuroscience, 10(2), 257–261 Vlassova, A., Donkin, C., & Pearson, J. (2014). Unconscious information changes decision accuracy but not confidence. Proceedings of the National Academy of Sciences, 111(45), 16214-16218. dundee.ac.uk Page 70