Perception Chapter 3 PDF

Perception Chapter 3 1 If our brains are information processors, where does the information come from? Everything we know comes from our senses. Our knowledge of the external world is completely dependent on the input provided by our sensory organs. I...

Perception Chapter 3 1 If our brains are information processors, where does the information come from? Everything we know comes from our senses. Our knowledge of the external world is completely dependent on the input provided by our sensory organs. If our senses provided extremely inaccurate information our concepts about the would deviate considerably from reality. This is the basis of the movie 'The Matrix'  Sensing the World Senses are natural adaptations that have evolved in response to the response selection pressures. Different organisms have developed different sensory capabilities because they fill different environmental niches. A frog feeds on flying insects; a male silkworm moth is sensitive to female sex-attractant odour; and we as human beings are sensitive to sound frequencies that represent the range of human voice.  Sensory Systems To represent the world, we must detect physical energy (a stimulus) from the environment and convert it into neural signals. Transduction is the transformation of physical stimuli (sights, sounds, smells) into neural impulses. Transducers are located in our sensory organs. Eyes, ears, skin...  Sensation & Perception Sensation: Encoding stimulus energy into neural impulses Extracting basic information about the stimulus Perception: Selecting, organizing and interpreting sensory signals The conscious experience of objects and scenes (relationships between objects) Overview: From Stimulus to Perception Energy contains information about the world e.g., wavelength of light is related to material properties of objects (we perceive this as colour) This information is usually limited, noisy & distorted Sensory organs capture and modify the energy Receptors transduce energy into a neural response Sensory nerves transmit the coded activity to the central nervous system and then to the brain  Overview: From Stimulus to Perception Thalamus processes and relays the neural response Relayed to specialized areas of the cortex Information is processed in parallel by a number of brain areas Perception of the world is created...  Perception is... A constructive processes that interprets information from our senses to allow us to produce an internal representation of the world sufficient for us to recognize and interact with our environment The representations are NOT exact copies of the world Perceptions are based: On estimates based on sensory signals On information from our experience and expectations Perception is a process similar to problem solving  Challenges to Perception Inverse Projection Problem Refers to the task of determining the object responsible for a particular image on the retina The retinal image is ambiguous many different object configurations could generate any retinal image  Challenges to Perception Information contained in the retinal image is limited Parts of objects can be hidden The image can be blurred or distorted e.g., smoke, glare, heat waves, etc.  Challenges to Perception Objects look different from different viewpoints The retinal image cast by an object is dependent on the orientation of the object relative to the perceiver Nevertheless, people are adept at identifying objects even when the orientation is novel to them (viewpoint invariance) This is something that artificial vision systems struggle with  Illusions Human perceptual systems are far from perfect and are susceptible to illusions  Influences on Perception Bottom-up processing: Perception based on incoming signal from the senses Energy registering on receptors Top-down processing: Perception based on information already stored in the brain Person’s knowledge, experience, expectations  Approaches to Understand Perception Direct perception theories Bottom-up processing Perception comes from stimuli in the environment Parts are identified and put together, and then recognition occurs Constructive perception theories Top-down processing People actively construct perceptions using information based on expectations Top-Down Processing  Constructive Perspective: Top-Down Processing Top-down processing affects perception by: Making inferences based on context Using knowledge obtained by previous experience Top-down processing occurs quickly and automatically Top-down effects are combined with bottom-up information from the senses to construct our perceptions Within the visual system sensory information from an object (green arrows) interacts with sensory information from the surrounding objects (blue) and feedback signals (red) based on prior knowledge and expectations. “Multiple personalities of a blob.” Top-Down Processing and Speech Unlike written language, in spoken language words are not neatly separated by gaps Often there is NO pause between words Top-Down Processing and Speech Speech segmentation: The process of distinguishing words in the continuous flow of speech sounds Speech segmentation is aided by context and linguistic knowledge When we learn a language we learn the probabilities (statistical learning) of sounds in different parts of a word (beginning, middle, end) and the probability of one sound following another (transitional probabilities) Helmholtz’s Theory Of Unconscious Inference Some of our perceptions are the result of unconscious assumptions we make about the environment We use our knowledge to inform our perceptions We infer much of what we know about the world Inference process is unconscious automatic Likelihood principle: we perceive the world in the way that is “most likely” based on our past experiences Based on your experience, which is a more likely interpretation of a? b or c?  Gestalt Laws of Perceptual Organization Gestalt proposed a series of “laws of perceptual organization” Rules that determine how sensory elements are grouped and organized to form objects  Gestalt Laws of Perceptual Organization Law of good continuation: Lines tend to be seen as following the smoothest path Objects overlapped by other objects are perceived as continuing behind the overlapping object  Gestalt Laws of Perceptual Organization Law of Good Figure (pragnanz): Every stimulus pattern is seen so the resulting structure is as simple as possible We see five circles, as in (a), not the more complex array of nine objects, as in (b). Gestalt Laws of Perceptual Organization Law of similarity: Similar things appear grouped together (a) This display can be perceived as either vertical columns or horizontal rows (b) more likely perceived as columns of blue dots alternating with columns of red dots, due to similarity of colour  Gestalt Laws of Perceptual Organization Law of familiarity: Things are more likely to form groups if the groups appear familiar or meaningful Law of proximity: Things near each other appear grouped together Law of common fate: Things moving in the same direction appear to be grouped together  Gestalt Laws of Perceptual Organization Gestalt laws often provide accurate information about properties of the environment The “laws” predict perceptions that correspond to what usually happens in the environment Gestalt Laws reflect: The evolutionary history of humanity Contemporary researchers think that the personal experience of individuals also plays a role  Gestalt Laws of Perceptual Organization Gestalt Laws do not always provide accurate perceptions e.g., a “coil of rope” may actually be a pile of short pieces of rope Gestalt “Laws” are not actually laws They are heuristics  Heuristics Heuristic: Provides best-guess solution to a problem “rule of thumb” Based on experience Fast Often correct: provides answers that have usually been correct in the past  Algorithms Algorithm: Procedure guaranteed to solve a problem (assuming all the necessary information is available) A set of well-defined computational steps Often Slow Definite result  Environmental Regularities Many aspects of the environment occur frequently and/or in predictable times and places This information can be used by the brain to improve perception Human perception is maximally sensitive to stimuli that occur frequently in the environment Environmental Regularities Oblique effect: People can perceive verticals and horizontals more easily than other line orientations  Environmental Regularities Light-from-above heuristic: Light comes from above Is usually the case in the environment We perceive shadows as specific information about depth and distance c  Semantic Regularities Semantic Regularities: Characteristics associated with the function and location of objects in different types of scenes Based on an individuals prior experiences with objects and locations  Palmer (1975) Presented a scene followed by a very brief presentation of a target object Participants attempted to identify the target object The target object was: An object that would be likely to be found in the scene An object that looked similar to an object that would be likely to be found in the scene An object that would not be found in the scene and was not similar to an scene consistent object  Palmer (1975) Semantic knowledge about context improved perception of the briefly present target. (a) a scene consistent object: ~ 80% accurate (b) scene inconsistent but visually similar: ~40% (c) scene inconsistent and visually dissimilar: ~40% Bayesian Style Inferences in Perception Helmholtz proposed that the brain uses the likelihood principle to resolve ambiguous sensory input But how does the brain determine how likely the various possibilities are? One way would be to apply Bayesian style inferences Bayesian Style Inferences in Perception Bayesian Inference: a statistical method for determining the probability of various possibilities Conclusions are based on both: Prior probability: the estimated probability of an event occurring in general Some events are likely; others are unlikely Likelihood given the current evidence (e.g. Sensory signal) How likely are the various possibilities to generate the observed evidence Bayesian Inference Bayesian Style Inferences in Perception Prior probabilities and likelihoods of objects and object configurations are learned through experience The inference procedure would be unconscious and automatic Bayesian inference has been successfully applied in the field of computer vision This indicates that this technique could possibly provide a basis for human vision Neuronal Codes and the Environment  Neurons and the Environment Our perceptual abilities are made possible by populations of neurons tuned to respond to frequently occurring stimuli The sensitivity characteristics of our neurons have be shaped by Natural Selection to provide perceptual abilities that improved the survival of our ancestors However, learning can also shape the response characteristics of neurons  Experience-Dependent Plasticity Experience-Dependent Plasticity: Changes in neuronal functioning due to experience Experiments have shown that neurons become more strongly tuned to aspects of the environment that occur frequently This effect is strongest during maturation but can also occur during adulthood  Magnitude of brain responses to faces and Greebles (a) before and (b) after Greeble training. The coloured areas Greeble stimuli used by Gauthier. Participants indicate brain activity. The FFA is located were trained to name within the white squares. each different Greeble.  Perception and Action Research has revealed that separate brain systems are responsible for conscious perception and for controlling motor action What & Where (Ungerleider & Mishkin, 1982) What pathway (ventral pathway): identification of objects perception and conscious awareness Where pathway (dorsal pathway): identifying an object’s location controlling & coordinating motor actions with objects  What & Where: Using Dissociation Logic To understand a complex system, you can logically deduce conclusions from “malfunctions” If damage to different areas of the brain cause very different deficits We can determine which areas are necessary for a specific function Brain Ablation method allows scientists to damage specific areas of otherwise normal brains (usually monkeys or cats) Controlled damage allows for clear conclusions to be drawn What & Where: Using Dissociation Logic Single dissociation: One function is lost, another remains Indicates different mechanisms are responsible for the functions Functions may or may not be independent Example: Monkey A has damage to temporal lobe. This monkey is no longer able to identify objects (what) but can still identify locations (where) Therefore, what and where rely on different mechanisms, although they may not be independent of one another What & Where: Using Dissociation Logic Double dissociation: Requires two individuals with different damage and opposite deficits Indicates different and independent mechanisms are responsible for the functions Example: Monkey A with temporal lobe damage has intact “where” but impaired “what” Monkey B with parietal lobe damage has intact “what” but impaired “where” Therefore, what and where streams must have different AND independent mechanisms The two types of discrimination tasks used by Ungerleider & Mishkin. Object discrimination: Pick the correct shape. Lesioning the temporal lobe (purple shaded area) makes this task difficult. Landmark discrimination: Pick the food well closer to the cylinder. Lesioning the parietal lobe makes this task difficult.  Neuropsychological Dissociations Hypothetical Example: (a) Alice can’t name objects but can accurately reach for them (b) Bert can name objects, but has trouble reaching for them. This illustrates a double dissociation. People with corresponding neurological problems have been found. Patient DF: Temporal Lobe Damage