Is visual perception influenced by nonvisual factors Dr

Is visual perception influenced by non-visual factors? Dr David Field Laura Inman

Spring term miniproject • How is this different from an Autumn miniproject? – You do not collect data • therefore there is much less emphasis on the Results section in the report – But you spend much more time reading articles and designing the study than you did for the Autumn miniproject • this is reflected in an increased emphasis on the Introduction and Method sections of the report

Divide into four groups • Groups should have 4 (or 5) members • Have you formed groups with people that you live with or know well? • To help you think about the role you play in groups, which you have been exploring in the Careers Skills module. . . • Each group has as number – all emails you send must include your group number in the subject line

What is the purpose of the visual system?

How does the visual system achieve its purpose? • What does it need to represent? • What is the conscious experience of having a visual system like? • How about the visual system of a house fly?

Embodied perception – Dennis Proffitt • Creating an accurate 3 D model of the world “in the brain” is one way to achieve the functions we believe vision needs to fulfil – If a system was designed (like artificial computer vision, or human vision if you believe in Genesis) then it might include a “vision module” of this type – but it is hard to envisage natural selection pressures operating over millions of years to create a system like that – natural selection is a blind force, operating in very small steps, and it does not aim to “design” anything – Rather, it builds systems up in a piecemeal way – Parallel coevolution of different parts of the brain probably produced a visual system that is not “modular”, or isolated from the purposes of other non-visual parts of the brain

Natural selection and the visual system • If an animal has a visual system trait that makes it more likely to reproduce than others in its species then that trait is likely to spread in the population – e. g. ability to see in low light levels • If an animal has a visual system trait that makes it less likely to reproduce than others in its species then that trait is likely to rapidly die out in the population – e. g. colour blindness that leads a herbivore to eat poisonous red berries

Evolution of the visual system • If due to a genetic mutation, an animal had a visual system trait that made it able to predict the motion of fast moving objects and so intercept them, then this might be a helpful adaptation – but this, like all visual system “purposes” can be achieved a variety of different computational algorithms, most of which do not require a 3 D reconstruction of the world “in the brain” • Because there is no sub-purpose of vision that absolutely requires an accurate 3 D representation of the world, it is unrealistic to hypothesize that selection pressure ever existed to produce such a system

Embodied perception proposal • Natural selection pressures have existed that favoured animals with visual systems which – make them energy efficient – reduce the chances of them choosing courses of action that would injure the animal – “A principal purpose of perception is to defend people from having to think” D Proffitt (my italics)

Embodied perception proposal – “A principal purpose of perception is to defend people from having to think” D Proffitt (my italics) • If dangerous, risky, or energetically costly environmental features are perceptually accentuated in some way and • features of the environment that imply high gains are perceptually accentuated in some other way then – the actor might make optimal and efficient choices rapidly and without occupying valuable cognitive resources – e. g. you don’t attempt to jump a chasm if it looks too wide: no need to think about it

Why does theory predict that nonvisual factors influence vision? • The answer is that visually based decision making processes need to take into account non-visual factors such as the action capabilities of the perceiver – If you perceive an un-jumpable chasm as much wider than a jumpable one that is only really a bit narrower, then this will result in adaptive behavioural choices, without spending time to think about each chasm encountered – But the implication of this is that the perceived width of the chasm is being influenced by a non-visual property (danger) as well as the retinal images that provide information about the actual physical width – Therefore we can predict that a tall person and a short person will perceive the same chasm as being of different widths

Embodied perception • Proffitt wishes to keep the embodied perception theory grounded in evolutionary theory and – therefore argues that non-visual factors should only cause biases in visual perception if those biases serve to make the actor better adapted to their environment – this is a part of theory you might choose to test as part of your miniproject

Evidence: slant perception • Participants stand at the base of a small hill and are asked to judge the slant of the hill in 3 ways – verbal estimate anchored at 0 and 90 degrees – visual matching by adjusting an angle on a disc – haptic matching (without looking at hand)

Verbal and visual judgments overestimate the slope Proffitt, Bhalla, Gossweiler, & Midgett (1995)

Verbal and visual judgments • 5 degree hills are overestimated by a factor of 4 – 10 degree hills by a factor of 3 – 30 degree hills by a factor of 1. 3 • Apparent slant is a decelerating function of actual slant – why should explicit awareness be biased in this way? Could this be useful and adaptive? – Clue: the maximum slope that can be ascended bipedally is about 30 degrees • The data show that humans can discriminate a 5 deg slope from a 6 deg slope – but 55 and 56 deg slopes would look the same – perceptual sensitivity is concentrated in the range where it is useful – the difference in energetic costs between climbing 5 and 6 deg slopes is large

Evidence: judging slope from the top of the hill • Judging the slope of a hill from the top instead of the bottom produces similar visual and verbal estimates unless…. – the real slope is > 25 deg, in which case estimates are greater when made from the top. Why? • (haptic measure still unbiased) • This difference cannot be due to increased energetic costs, but could be due to fear of falling and injury – avoiding injury is clearly adaptive…. . – http: //www. youtube. com/watch? v=KOy. QBSMe. Ih. M

Predictions from “energy efficiency” • If Proffitt is right, what non-visual manipulations should influence slope judgments? – wearing a heavy backpack increases the energetic costs that would be incurred by climbing the hill – being physically fit reduces energetic costs by increasing fuel efficiency • also, total available energy reserves are greater – being old increases energetic costs by reducing fuel efficiency • also, total energy reserves are lower – being tired increases energetic costs of climbing by reducing fuel efficiency • also, total energy reserves are lower

Data • All of these predictions are supported – hills appear steeper when the energetic costs of climbing them are higher relative to the energetic potential of the actor – although the haptic measure is never influenced by manipulations of energy cost

Scepticism • Most vision scientists think that this theory is crazy • They adhere to some version of the view that vision achieves it’s goals by creating an approximately accurate representation of the real world that can then be accessed by the motor system • They point to weaknesses in the experimental evidence for “embodied perception”

Scepticism • Proffitt & colleagues write about the slope judgment data as if the verbal, visual, and haptic measures of slope are direct measurements of perception • The problem is that overt behavioural responses are not pure measures of perception • Perception -> calibration and other processes -> overt behavioural responses • Sceptics argue that biases in slope judgments are a product of calibration and other intervening processes – therefore the biases are caused by post-perceptual processes

Durgin (2009) proposes that slope findings are explained by demand characteristics • 3 experimental groups – no backpack – standard backpack condition (like Proffitt) – electrode backpack condition (deceived into thinking equipment in backpack was needed so that electrodes could measure ankle muscles) • slope stimulus was a sturdy wooden ramp • verbal slope judgments • after slope judgment, asked why they thought they were wearing the backpack, and in electrode group, did they really believe backpack recorded data

Slope judgment influenced by cognition, not presence vs. absence of backpack • Among the standard backpack group, 12 of 13 guessed that the backpack was intended to alter their slope perception – 5 of the 13 said that it had influenced their judgments, and these 5 gave the steepest slope judgments in the experiment • 75% of those in the electrode backpack condition were successfully deceived into believing that the backpack was needed for measurement – their slope judgments were the same as the no backpack group

Proffitt (2009) – reply to Durgin

Debate / evidence / your project • There is other more evidence for the embodied perception theory – sports performance – reaching and grasping • I will put some starter papers on the above topics on my web page • We don’t want 18 miniprojects about hill slope judgments • You can use any method you like to address the question – brain imaging – animal experiments • You can introduce a new independent variable, e. g. – Fear of heights / claustrophobia – Do these phobia produce perceptual biases – Are these things that happen when this general mechanism that is normally adaptive goes wrong? • You should attempt to test the proposal that perception is influenced by non-visual factors, but only by non-visual factors that influence how well the actor is adapted to their environment

1) Proffitt, D. R. (2006). Embodied Perception and the Economy of Action. Perspectives on Psychological Science, 1(2), 110 -122. • Essential reading – make sure you read this one 2) Canal-Bruland, R. , & van der Kamp, J. (2009). Action goals influence action-specific perception. Psychon Bull Rev, 16(6), 1100 -1105. • uses children as participants 3) Durgin, F. H. , Baird, J. A. , Greenburg, M. , Russell, R. , Shaughnessy, K. , & Waymouth, S. (2009). Who is being deceived? The experimental demands of wearing a backpack. Psychon Bull Rev, 16(5), 964 -969. 4) Murray, S. O. , Boyaci, H. , & Kersten, D. (2006). The representation of perceived angular size in human primary visual cortex. Nat Neurosci, 9(3), 429 -434. 5) Proffitt, D. R. (2009). Affordances matter in geographical slant perception. Psychon Bull Rev, 16(5), 970 -972. • this is Proffitt’s reply to Durgin et al. (2009) 6) Wesp, R. , Cichello, P. , Gracia, E. B. , & Davis, K. (2004). Observing and engaging in purposeful actions with objects influences estimates of their size. Percept Psychophys, 66(8), 1261 -1267. 7) Witt, J. K. , & Dorsch, T. E. (2009). Kicking to bigger uprights: field goal kicking performance influences perceived size. Perception, 38(9), 1328 -1340. • the strongest evidence yet to emerge from the sports performance paradigm 8) Witt, J. K. , Linkenauger, S. A. , Bakdash, J. Z. , & Proffitt, D. R. (2008). Putting to a bigger hole: golf performance relates to perceived size. Psychon Bull Rev, 15(3), 581 -585. 9) Witt, J. K. , & Proffitt, D. R. (2005). See the ball, hit the ball. Psychol Sci, 16(12), 937 -938. 10) Bhalla, M. , & Proffitt, D. R. (1999). Visual-motor recalibration in geographical slant perception. J Exp Psychol Hum Percept Perform, 25(4), 1076 -1096. • • The original detailed report of the backpack studies - also described in reference 1 above Woods, A. J. , Philbeck, J. W. , & Danoff, J. V. (2009). The various perceptions of distance: an alternative view of how effort affects distance judgments. J Exp Psychol Hum Percept Perform, 35(4), 1104 -1117. – this one is only available as hard copy in the library. I don’t have a copy, and if you want to read it in detail I’d be grateful if you’d make me a copy too.

Week 2 • In summary, you are designing a study to test theory that – non visual factors influence visual perception – this in turn influences decisions about how to interact with the world in a way that optimises • energy efficiency and avoidance of danger • and maybe other things too? – but the effect of non visual factors should be limited to cases where there is an adaptive advantage to the bias • or at least that there was an advantage at some point in the evolutionary history of the species

Week 2 • Some researchers dispute the interpretation of the data suggesting that non visual factors influence perception – so you might design a study that attempts to resolve that question • there are other cases where is it is generally accepted that perceived dimensions of objects are quite different from the dimensions implied by their retinal images • These cases make theory seem more plausible. Can you think of the examples?

Size “illusion”

Week 2 • Or you might sidestep the dispute over the evidence and focus on predictions made by the second part of theory • that the perceptual biases should be adaptive ones

Guidance notes for report writing • These guidelines are in addition to those sent you by Tony Watkins – if there is any contradiction then follow these • Results section – No reporting of actual statistics, such as F values and p values, and degrees of freedom. • it does not make sense to do this when there is no data – But you should include graphs illustrating possible patterns of results • typically, there are two possible patterns, and you should not report more than three alternative patterns • this will require you to make up some summary statistics (e. g. means) in excel

Guidance notes for report writing • Confounds – This is an exercise in designing and planning scientific research – One of the most important aspects of planning science is to consider potential confounds, and how to avoid them – A confound is something that covaries with the IV, and which might offer an alternative explanation of the effects of the IV • e. g. hypothesize that men are better than women at basketball because they have superior athletic ability • one confound is that men are taller than women, and height is an advantage in basketball • to avoid the confound and test the hyp properly you need to compare men and women of the same height

Guidance notes for report writing • Confounds – This is an exercise in designing and planning scientific research – One of the most important aspects of planning science is to consider potential confounds, and how to avoid them – A confound is something that covaries with the IV, and which might offer an alternative explanation of the effects of the IV • e. g. hypothesize that hills appear visually steeper to people wearing backpacks because of increased potential energy costs of climbing the hill • confound arises if people wearing backpacks guess why they are being asked to wear them in the experiment • increased slope judgment could be due to either or both of the above factors

Guidance notes for report writing • Confounds – Because consideration of confounds is so important when designing experiments, I want you to include a “Confounds” sub heading in your method section, where you discuss potential confounds in your experiment, and how your design attempts to control for them. • note that it is not always possible to avoid all confounds!

Guidance notes for report writing • Confounds – Because confounds are so important when designing experiments, I want you to include a “Confounds” heading in your method section, where you discuss potential confounds in your experiment, and how your design attempts to control for them. • note that it is not always possible to avoid all confounds! – This section can be incorporated into the Design section of the Method if you wish

Guidance notes for report writing • Finance – You need to provide some estimates of the costs of carrying out your project in the method section • there are is no upper limit, but… – In the Discussion section you can consider whether your project offers value for money • note that value for money is not the same as low cost

Guidance notes for report writing • Literature – last year, students often included either too much or too little detail about published studies that they mentioned • A heuristic, or rule of thumb for reporting literature – If your study will use a similar methodology to the study you are describing, and depends on the study for it’s theoretical basis, then describe the study in detail, perhaps with some critical comments • explain it’s method and findings, and then say how your method will be different and why – If the study you are describing is part of the general background, then limit yourself to describing the “take home message” from the study

Is the softball result predicted by theory? • Proffitt wishes to keep the embodied perception theory grounded in evolutionary theory and – therefore argues that non-visual factors should only cause biases in visual perception if those biases serve to make the actor better adapted to their environment • It is hard to see the adaptive value of increasing the apparent size of a ball when you are doing well – Witt speculates that these effects are due to historical links between sport and hunting – A hunter who is not very good at throwing his spear will see the target as smaller, encouraging him to get closer to the target – A hunter who can throw his spear accurately will see the target as bigger, and will be encouraged to throw his spear from further away, thereby gaining the advantage of a lower risk of detection by the target animal