L1.2 Visual WM 2024 Students PDF

PSYC327: Advanced topics in Cognition Associate Professor Simone Favelle Visual Memory: Visual Working Memory & STM SIMONE FAVELLE PSYC327 - ADVANCED TOPICS IN COGNITION What happens after iconic memory? Iconic memory is good for temporary storage… But we do need to store informati...

PSYC327: Advanced topics in Cognition Associate Professor Simone Favelle Visual Memory: Visual Working Memory & STM SIMONE FAVELLE PSYC327 - ADVANCED TOPICS IN COGNITION What happens after iconic memory? Iconic memory is good for temporary storage… But we do need to store information for longer – Is that object the same one I saw 10s ago? What happens after iconic memory? Three broad components of visual memory (~ multi store model): – Iconic memory – Visual STM/WM – Visual LTM Lecture outline What is Visual STM/WM? distinct from iconic memory Visual WM distinct from verbal WM Visual and spatial information in STM What is the content of Visual WM? What might influence visual WM capacity? STM vs WM Often used interchangeably, but Short term memory (STM) is the capacity for holding, but not manipulating information Working memory (WM) is a theoretical framework of structures and processes (eg, rehearsal, attention) used for the temporary storage and manipulation of information, of which STM is just one part. PROBLEM: Visual memory was being investigated in terms of familiar items with labels and semantic information attached Question: How do we investigate visual information alone in memory? Answer?? Phillips (1974) Observers were presented with pairs of matrices with a random half of the cells black and the other half white Complexity of matrices varied – (4x4, 5x5, 6x6 & 8x8) Observers had to make a same/different judgment Phillips (1974) Experiment 1: What happens to visual memory after the 1 s of iconic memory? The interval between the pairs of displays (ISI) was varied between 0 - 9s.  Sequential Matching Task + 1s 0–9s Same/different judgment PRESENT TEST 4x4 6x6  Demo: are the matrices the same or different? +  How do you think the size+ of the matrices affected performance? What about the ISI? Let’s look at the results…  4x4 100 6x6 90 8x8 80 % Correct 70 60 50 0.2 1 3 9 ISI (s) Phillips (1974) Regardless of complexity, performance at intervals < 0.5s was near perfect as expected from iconic memory Performance declined as intervals increased and relied more on complexity Suggested that this might be Visual STM Phillips (1974) Experiment 2: What are the properties of visual STM? Iconic memory thought to be based on absolute retinal coordinates Visual STM/WM might use object-based coding Can iconic memory be disrupted by changing the absolute location of the matrices? – Stimuli presented at same location or jittered  Pattern complexity was varied + The duration of ISI was varied Pattern location was jittered Same/different judgement  5x5, same 100 5x5, jittered 8x8, same 90 8x8, jittered 80 % Correct 70 60 50 0.2 1 3 9 ISI (ms) Jitter curves were essentially flat No input from iconic memory, only STM contributing to performance Phillips (1974) Experiment 3: Masking Iconic memory (and not visual STM) should be sensitive to masking Can iconic memory be disrupted by masking the matrices? – Stimuli presented at same location but with checkerboard mask or blank during ISI  Pattern complexity was varied + The duration of ISI was varied and masked Same/different judgement  5x5, same 100 5x5, masked 8x8, same 90 % Correct 8x8, masked 80 70 60 50 0.2 1 3 9 ISI (ms) Similar to jitter results - curves were flat No input from iconic memory, only STM contributing to performance Phillips (1974) Conclusion – The two parts of the curve reflected different memory stores: – Iconic memory (Early part of the curve) brief, large capacity, information is tied to spatial position – Visual short-term memory (Later part of the curve) long lived, limited capacity, information not tied to spatial position Visual Working Memory (VWM) What is VWM? Is it distinct from verbal and spatial WM? VSSP? Not clear how or if there is rehearsal of visual information Neuropsychological Evidence Patients KF (Warrington & Shallice), and PV (Vallar & Baddeley, 1984). Severely impaired verbal short-term storage (e.g. digit span) with auditory presentation. Digit span for visual presentation much better. Points to separate STM systems for visual and for verbal material Logie (1986) Subjects rehearsed words by forming images (image group) or by repeating the words (rote group) – Beach, pillow, apple, car, shoe…. Distracted subjects with either irrelevant patterns (matrices) or words  Patterns disrupted image group and not rote group (some disruption) – words disrupted rote group and not image group Logie (1986) Experimental double dissociation Selective interference within modalities suggests specific stores for visual and articulatory information Visual Patterns test Della Sala, et al (1997) Measure of visual STM capacity Memorize b&w matrices of increasing complexity and reproduce them Maximum complexity = VSTM capacity “Pure” Visual STM Dissociate visual from spatial STM Della Sala et al (1999) Sample of brain-damaged adults Patient Group A Patient Group B (n = 2) (n = 1) Spatial task (Corsi blocks task) ✗ ✓ Visual task (Visual Patterns task) ✓ ✗ VWM – Features or Objects? Visual information about objects can be stored in two general forms: – Individual features: no connection to other information (e.g., colour, shape, location). – Bound object features: grouped together in “chunks” (e.g., colour-location pairs, object files). Luck & Vogel (1997) Developed a change detection procedure for isolating VWM Participants made a same–different judgment after seeing two temporally separated displays of multiple visual objects Half of the displays identical, half differed by a single feature Now you try… Get ready to record your response Look at the fixation cross (it will remain on screen) First display flashes quickly and second remains a little longer Respond same (s) or different (d) 6 trials, brief pause in between  VWM is object-based (does not differ across number of features) Capacity of about 4 objects Objects or Locations? Lee & Chun (2001) Is VWM capacity determined by objects or spatial locations? Change detection task – number of locations was varied while the numbers of objects and features were held constant Lee & Chun (2001) Lee & Chun (2001) Found no difference between together and separate conditions The number of spatial locations does NOT influence VWM capacity Does the complexity of the object matter? Alvarez & Cavanagh (2004); Eng, Chen & Jiang (2005) Are coloured squares equivalent to faces? Change detection task – to determine capacity Visual search tasks – to determine informational load (object complexity) Eng, Chen & Jiang (2005)  Yes, VWM is sensitive to object complexity Capacity decreases as object become more complex Visual WM… is best measured with novel stimuli is distinct from iconic memory (shown by Phillips, 1974) can be dissociated from both verbal WM and spatial STM capacity depends on number of objects (not features), object complexity, not based on object location What is the role of VWM? Tracking objects across locations (iconic memory is disrupted by jittering) Guide eye movements, gaze corrections