Memory

• Neisser (1967) - Cognition involves “all processes by which the sensory input is transformed, reduced, elaborated, stored, recovered, and used."
• Gibson’s Ecological Theory - perception = sensation - implies that no cognition is required and that we perceive things through object affordance, meaning we recognise objects automatically
• • Gregory’s Constructive Theory - perception is indirect and hypothesis testing is used to interpret the input. He would argue that cognition starts with perception.
• Modal Model of Memory by Atkinson and Shiffrin (1968) - describes how environmental stimuli is processed through the memory modules of sensory, short-term, and long-term memory.
• Criticisms of Model Model - it cannot explain retrieval through a linear flow of information and neuropsychological evidence struggles to support it.
• Support for Model Model - it acknowledges distinct types of memory stores and there is empirical evidence for the existence of its components
• Illusory motion works through sensory memory storing the incoming information for long enough for to bridge the gap between the separate static images. If the gap is too long, it may exceed the sensory memory duration capacity, and motion may not be perceived
• Sperling’s (1960) whole report showed that participants who were shown 12 letters for 50ms were only able to recall ~4 letters, showing how the information had decayed during the retention period.
• Sperling’s (1960) partial report had participants only recall one of the 3 lines of 4 letters, signalled with an auditory cue, with a variable delay from 0-1s. No delay = ~3/4 of the letters. Short delay (300ms) = ~2/4. Long delay (1000ms) = ~1/4. This shows how sensory memory has a retention time of ~1 second with a total capacity ~9 letters
• Crowder (1971) - serial auditory presentation of 9 digits with a verbal and non-verbal recall cue. Found primacy and recency effects in recall with the non-verbal cue, and found that recency effect is attenuated when the cue is a spoken word. Shows that echoic memory seems to be modality-specific and rapidly overwritten
• Primary and recency effects (serial order effects) also shown by Glanzer and Cuntiz (1966)
• Sensory memory seems to be modality specific (Sperling, 1960) and short lived (Crowder, 1971), with <1s for iconic and ~2s for echoic memory. Sensory information is ultimately lost once memory has decayed or is replaced by new information
• Attention is the mechanism by which we select information that is relevant to our current behaviour. Information that is selected, or attended to, gains access to higher cognitive functions, such as memory. Sensory memory is often understood as the interface between perception and memory.
• STM encodes information phonologically and holds a small amount of information for a short period of time
• Conrad and Hull (1964) - tested participants with a serial recall of consonants, which had to be recalled in the correct sequence to test their memory span. The two conditions were letters that were phonologically similar or dissimilar. They found error rates were higher for phonologically similar letters, showing that the STM seems to store information phonologically.
• Miller (1956) - reviewed other research, finding STM capacity limited to 7 (+/-2) items, with chunking (rhythm/category) increasing the number of information we can retain
• Glanzer and Cunitz (1966) - free recall of one syllable nouns either immediately, or after a delay of 10 or 30s. Found that the first (primacy) and last (recency) words are recalled better than the intermediate words when items are recalled immediately. Delay eliminates recency effects, meaning that STM traces decay between 10-30 seconds and that primacy effects reflect LTM performance through sub-vocal repetition.
• The phonological loop has two sub-components; a phonological short term store with limited capacity and duration, and an articulatory control process with sub-vocal rehearsal and translation of visual information.
• The phonological loop can explain the 3 prominent effects of verbal STM; the phonological similarity effect, the word length effect, and the irrelevant sound effect.
• Baddeley (1966) - participants had to serially recall auditory presented word lists. There were 4 conditions; phonologically similar and distinct words, and semantically similar and distinct words. Found that recall accuracy was impaired for phonologically similar words and performance is similar for semantically similar and dissimilar words. This shows how the Loop stores information phonologically.
• Word Length Effect - Baddeley et al. (1975) - participants had to serially recall visually presented words of 1, 2, 3, 4, or 5 syllables. Found that the longer it takes to read the list of words, the lower the accuracy of recall, finding that people can remember as many words as they can say in ~2.5s. However, there are other factors such as decay, interference, and volume.
• Irrelevant Sound Effect - Colle and Welsh (1976) - participants had to serially recall auditory presented 8-letter lists, presented to them with noise (foreign language words) and without noise. Found that error rates were higher with noise and shows primacy effect (LTM). Shows that phonological material is automatically encoded in the Phonological Loop.
• Phonological Similarity and Irrelevant Sound Effect are effects of the phonological short-term store, and the Word Length Effect is an effect of the articulatory control process
• Non-nameable visual material (Philips, 1974) - Change detection task with a memory display of a checkerboard pattern with 4x4, 6x6, or 8x8 cells, a retention period of 0-9s, and a test display with an identical pattern or one-cell change. Found ~100% accuracy with immediate recall, with accuracy decreasing over time and for larger matrices, meaning small amounts of non-nameable visual material is retained for a short period of time. This shows that there is a temporary short-term store with a limited capacity for non-nameable visual material
• Luck and Vogel (1997) - Change detection task with 1, 2, 3, 4, 8, or 12 coloured squared, a retention period of 900ms with an articulatory suppression task of counting to 3 or without, and a test display that was identical or with one colour changed. Found accuracy decreases from a set size of 3, regardless of articulatory suppression. Shows further evidence for a limited capacity (3) storage for non-vocalised visual material.
• STM has 2 separate stores, one for auditory and vocalised visual info, The Phonological Loop, and one for non-vocalised visual information
• Visual STM has 2 components: visual object memory (what it is) and spatial location memory (where it is) (Ungerleider and Mishkin, 1982)
• Separate auditory and visual STM stores shown through double dissociations (Patient KF, Shallice and Warrington, 1970; Patient MV, Carelsimo et al., 2001) and through neurotypical behaviour of dual task performance
• Working Memory (Baddeley and Hitch, 1974) - central executive (attentional controller), visuospatial sketchpad, and phonological loop
• Phonological Loop - appears to be involved in language acquisition - Patient PV (Baddeley et al., 1988) has a digit span of 2, but has a preserved visual STM, demonstrating a pure Phonological Loop deficit. PV and controls learnt 8 pairs of Italian/Italian words or Italian/Russian words with a spoken presentation. PV could not learn any Italian/Russian words.
• Phonological Loop and Language Acquisition - learning foreign language words is harder with articulatory suppression (Papagno, Valentine, and Baddeley, 1991) and for phonologically similar and longer words (Papagno and Vallar, 1992)
• Visuospatial Sketchpad - Shepard and Metzler (1971) conducted a mental rotation task, finding the greater the angle of rotation, the more time it takes to decide whether the two objects are the same, showing that the real world distance equates the mental distance, supporting the idea that there is storage to map a mental image
• Visuospatial Sketchpad - Kosslyn et al. (1978) found a linear relationship between the location distance on a map and the time taken to mentally travel between the locations
• Visuospatial Sketchpad - Milivojevic et al. (2003) mental folding task, finding a linear relationship between the number of folds and the time taken to mentally fold the shapes.
• Central Executive - Baddeley et al. (1991) found that performance in a dual task of visuospatial tracking and digit span was poorer in patients with Alzheimer's, suggesting attentional resources are unequally distrbuted.
• Criticisms of Working Memory:

  • Fails to explain semantic chunking - can retain up to 16 words (Vallar and Baddeley, 1987) and semantic chunking seen in amnesic patient, meaning it exists in the absence of LTM (Baddeley and Wilson (2002)
  • Fails to explain visual similarity effects in verbal recall - small advantage for visually dissimilar but acoustically similar words (Logie et al., 2000)
• Episodic Buffer (Baddeley, 2000; 2012) - a storage system that holds chunks of multidimensional information via connections to the Sketchpad, Loop, and LTM. The capacity is 4 chunks of multi-dimentional information and loops of rehearsal and retrieval generates and activates long term information.
• Function of WM - Baddeley (1968) found a concurrent STM task slows down reasoning speed, but doesn't prevent it, finding that information retention and information manipulation work in parallel
• Learning - Total Time Hypothesis (Ebbinghaus, 1885)

  • More learning repetitions of meaningless and nonsense syllables on day 1 meant less time was required to re-learn the lists on day 2
• Learning is not just time invested:

  • Ericsson et al. (1993) - 'good' students and 'best' students practiced the same amount of time per week
  • Extensive practice enhances performance, but eventually plateaus. Deliberate practice is needed to improve