Psychology of Perception
Psychology of Perception 222
Popular in Psychology of Perception
verified elite notetaker
Popular in Psychology
This 5 page Class Notes was uploaded by Sarah Kincaid on Thursday September 29, 2016. The Class Notes belongs to 222 at Boston University taught by Rucci in Winter 2016. Since its upload, it has received 12 views. For similar materials see Psychology of Perception in Psychology at Boston University.
Reviews for Psychology of Perception
Report this Material
What is Karma?
Karma is the currency of StudySoup.
You can buy or earn more Karma at anytime and redeem it for class notes, study guides, flashcards, and more!
Date Created: 09/29/16
Week Four If given hits probability, know miss probability. Stimulus (presence of stimulus) If given correct rejection probability, know false alarm Present Absent probability. (absence of stimulus) Response Yes Hit False Each pair goes hand in hand. alarm d' = how far away two distributions are from each other No Miss Correct d' defines how different distributions for rejection stimulus present and stimulus absent d' = 0 means that the distributions are nearly on top of each other, difficult to discriminate, no Signal detection theory (SDT): -distinguishing signal from noise sensitivity -Helps with decision making under d' = 1 means moderate sensitivity, d' greater or conditions of uncertainty equal to 1 is relatively easy to discriminate between Ex. Radiologists trained to tell d' = 4 means high sensitivity whether cancer or just noise on mammogram. Ex. Air traffic controllers – is it a flock of birds (noise) or a plane? Note: The farther you get away from the population mean, the less likely it is the signal. Ex. WW2 – is it an enemy sub or a whale (noise)? – risk, giving away position. Sensitivity ease with which an observe can differentiate between presence and absence of stimulus or the difference between two stimuli separation between distribution of response to noise alone and signal plus noise Blue line = noise, black line = signal (in pic above) Pr(Hit) + Pr(Miss) = 1 Pr(FA) + Pr(CR) = 1 Lower Criterion = Liberal Criterion (e.g. mammograms) More hits More false alarms Willing to call more things “signal” Unbiased Area of miss = area of false alarm or near the intersection of the two distributions Week Four Higher criterion = conservative criterion (e.g. peanut allergies, only exclude foods you are positive have peanuts) fewer hits fewer false alarms not willing to raise false alarm Autokinesis: small point of light in dark environment appears to move Dot moving (stimulus 30'/s 48/50 0.96 (hits) present) Dot still (stimulus absent) 0 4/44 0.09 (false alarms) Figure 1. Dot only presented for one second - if presented longer, dot still would have eventually become 44/44 (1.00 false alarm because of autokinesis) Figure 2. Position criterion so that the area above the curve will be high for hits and low for misses Z score = how many standard deviations a raw score is above or below the population mean d' = z (Hits) - z (False Alarms) d' = 1.33 - (-1.75) = 3.08 (for above figure) Receiver Operating Characteristic Curves (ROC) graph hit rate as function of false alarm rate at Pr(Hit) = Pr(FA), you are guessing whether the signal is present or not o appears as a diagonal line as observer sensitivity increases, d’ increases, and the curve bows up/outwards toward upper left corner (100% hits, 0 % FAs) once d’ reaches 3, it does not make a difference how far apart the distributions are Week Four Figure 3. ROC curves The blue line (worthless) has a ~1:1 ratio of hits: false alarms. False alarms: 0.988, Hits: 1.000. The black line (good) has a better ratio. More hits and less false alarms. The grey line (excellent) has the best ratio. Lots of hits and few false alarms. *Note: Different radiologists may use different criterion - false alarm for cancer is bad because of worry and harsh treatment however the consequences of missing cancer are death so worth the low criterion Frequency Analsis (Fourier, Spectral) Experimental study of perception Gratings (horizontal and vertical stripes) Jean Baptiste Joseph Fourier, "Theorie analytique de la chaleur" (1822) Representation Fourier series periodic functions Fourier transform Non-periodic functions Sine Wave (sinusoidal functions) has 3 parameters: Amplitude (contrast) Frequency Phase In the figure to the right, the amplitude is the height of the oscillation from the position of equilibrium (x-axis). There are 2 periods in one se-1nd, thus the frequency is 2 Hz (or 2 s ). Phases will be explained on the next page. Week Four When there are 12 periods in 1 s, the frequency is 12 Hz. In this case, the period (T) is 1/12 s. Phases are measured in angular units (degrees or radians). The phase is how the sine wave is positioned along the axis. Phases are how much the sine wave has been shifted. G1: 0 T G2: ¼ T G3: ½ T Sounds: changes in pressure over time Tuning forks produce pure tones. Pure tones change pressure over time. When a wavelength is halved, the frequency doubles. Higher frequency & smaller amplitudes signal (closer to straight line) An alternative representation, Fourier/Frequency/Spectral Analysis If you shift the function more (time vs. amplitude), then the line will go up or down on the frequency vs phase graph. If you increase the frequency, the dot/line will shift right. Fourier components: gratings aka gradients - Add gratings together forms patterns and builds images (neuroscience) Week Four Fourier analysis - Sinusoidal function: amplitude, phase, frequency - Plot Amplitude vs frequency and phase vs frequency - sum of sinusoidal functions forms signal amplitude = contrast in images, amplitude is contrast of color (higher amplitude = sharper contrast) same frequency, different amplitude = same stripe width, different color contrast same amplitude, different frequency = same color contrast, different stripe width sharp transition = high frequency filtered out (filtering) low frequency filtered out = edgy look our vision prefers specific range of frequencies (some are low and appear big and blurry, some are high and appear tiny – our distance from them determines whether we can see them at lower contrast (amplitude = contrast in 2d)
Are you sure you want to buy this material for
You're already Subscribed!
Looks like you've already subscribed to StudySoup, you won't need to purchase another subscription to get this material. To access this material simply click 'View Full Document'