Session 1: When Expectations Meet Reality
Session 1: When Expectations Meet Reality
Overview
This session examines how prior knowledge and expectations modulate sensory processing and subjective experience. We'll begin with the Pepsi paradox, where Montague's 2004 fMRI study demonstrated differential neural responses when participants knew which brand they were drinking versus blind conditions, with brand knowledge activating prefrontal regions associated with memory and higher-order cognition. We'll review the predictive processing framework, which proposes that perception results from the brain's continuous generation of predictions about sensory input, with conscious experience reflecting the integration of bottom-up signals and top-down priors. Additional studies demonstrate this principle across modalities; odor labels (e.g., "cheddar cheese" vs "body odor") change experienced pleasantness of identical smells and corresponding brain activity in olfactory regions; placebo treatments produce measurable changes in neurotransmitter release and pain-related brain activation; packaging and brand presentation alter taste perception in controlled experiments. We'll discuss the methodological approaches researchers use to isolate expectation effects from sensory processing, including blinded designs, neural decoding, and computational modeling. This work challenges traditional models of perception as passive stimulus processing and raises questions about how to define and measure "veridical" subjective experience when top-down influences are ubiquitous.
Why Do Expectations Matter?
The Traditional View: Perception as Window to Reality
Historically, perception has been conceived as a relatively straightforward process: sensory organs detect physical properties of the world (light waves, sound waves, chemical molecules), transduce them into neural signals, and the brain reconstructs an internal representation of external reality. In this view, perception functions like a camera or recording device—faithfully capturing what's "out there" and delivering it to consciousness.
This model implies that subjective experience should be relatively uniform across individuals exposed to the same stimuli, and that perceptual judgments should be largely independent of factors like beliefs, expectations, or cultural background. Disagreements about perception would reflect differences in sensory acuity or attention rather than fundamental differences in what is experienced.
The Challenge: When Reality Meets Mind
However, decades of research reveal that perception is far more complex than passive reception. The brain doesn't just register sensory input—it actively interprets, predicts, and constructs experience based on prior knowledge. This raises fundamental questions about the relationship between objective reality and subjective experience:
- If two people receive identical sensory input but have different prior experiences, do they have the same subjective experience?
- Can we ever perceive "pure" reality without interpretation?
- Where does sensation end and cognition begin?
- What does it mean for perception to be "accurate" if it's always filtered through expectations?
Language and Cultural Shaping of Perception
Some of the most striking examples of expectation effects come from language and culture, which provide frameworks for organizing and interpreting sensory experience:
- Categorical perception: Once we learn categories (linguistic or otherwise), we perceive within-category differences as smaller and between-category differences as larger than they physically are. This affects multiple domains:
- Color: Russian speakers discriminate blues faster across their linguistic categories
- Phonemes (speech sounds): Studies show cross-linguistic differences:
- Werker & Tees (1984) - Infants lose ability to discriminate non-native phoneme contrasts by 10-12 months
- Goto (1971) - Japanese adults have difficulty with /r/-/l/ distinction that English speakers perceive categorically
- Kuhl (1991) - "Perceptual magnet effect" showing prototypical phonemes make nearby sounds seem more similar
- Facial expressions: Categorical perception of emotional expressions
- Cultural display rules: Culture shapes which emotions we attend to and how we interpret facial expressions.
- Perceptual expertise: Wine experts, musicians, and radiologists don't just know more—they literally perceive differently, seeing or tasting distinctions that novices cannot. This expertise reflects learned expectations about what to attend to and how to parse complex stimuli.
The Invisible Gorilla in the X-ray: Why? Because their memory-trained perceptual systems were honed to detect tumors—not gorillas.
Figure from Drew, Ho & Wolfe, 2014
🎥 Watch: Dr. Biyu He on Conscious Experience
Prior Knowledge as Interpretive Framework
What we know shapes what we perceive in multiple ways:
- Attention guidance: Knowledge directs where we look and what we notice. Experts scan visual scenes differently than novices.
- Disambiguation: When sensory input is ambiguous (which is often), prior knowledge determines which interpretation wins. The same sound can be heard as speech or noise depending on linguistic knowledge. Viewing ambiguous images (like the Necker cube or face-vase illusion), spontaneous fluctuations in brain activity—before the stimulus even appears—predict which interpretation people will perceive. This suggests that ongoing brain states, shaped by prior expectations and recent history, actively determine conscious experience even with identical sensory input.
How Paranoia Shapes Perception: Finn et al. (2018) demonstrated that individual differences in paranoid thinking predict how people interpret ambiguous social information. In their fMRI study, participants higher in paranoia showed heightened connectivity between threat-related brain regions when viewing the same ambiguous scenarios that low-paranoia individuals found benign. Prior beliefs about threat literally changed which neural interpretation won out, even with identical sensory input.
- Filling in: The brain uses prior knowledge to "fill in" missing information—we literally see things that aren't there based on what we expect to be there.
- Integration: Multiple sources of information (visual, auditory, semantic) are weighted and combined based on learned reliability, creating unified percepts.
Predictive Processing Framework
The Basic Idea
To understand how expectations shape experience, we need to examine two fundamentally different models of how perception works:
- Sensory information flows "bottom-up" from your eyes/ears/etc. → through processing stages → to your brain's higher areas
- Like climbing a ladder: raw input → simple features → complex features → recognition
- The brain is basically passive, waiting to see what comes in
- Your brain is constantly making predictions about what it expects to sense
- These predictions flow "top-down" from higher brain areas → back to sensory areas
- Only the errors between predictions and actual input flow upward
- The brain is actively trying to minimize surprise
A Simple Example
Imagine you're at home and hear your front door open:
Sound waves → ear → auditory processing → "that's a door" → "someone's home"
Predictive Coding Model:
- Your brain already predicts: "My partner comes home around 6pm"
- When you hear the door at 6pm, there's no prediction error (expected sound matches reality)
- Very little neural activity needed! The prediction was right.
- BUT if that door sound happens at 3am, there's a huge prediction error
- Lots of neural activity as your brain updates its model and predicts what's happening
Key Principles
1. Hierarchical Predictions
- High-level areas predict what mid-level areas should be doing
- Mid-level areas predict what low-level areas should be doing
- All the way down to predicting individual photons hitting your retina
2. Prediction Errors Propagate Upward
- When predictions are wrong, the error signal goes up the hierarchy
- Higher areas update their models
- Send new predictions back down
3. Precision Weighting
- The brain tracks how reliable different signals are
- Attention = increasing the precision of certain prediction errors
- "Pay attention to this" means "weight these prediction errors more heavily"
Concrete Neural Evidence
Visual Cortex Studies:
- When you see a predictable stimulus (like a repeating pattern), neural activity in V1 (primary visual cortex) decreases over time
- The brain is getting better at predicting it, so less prediction error
- This is the opposite of what you'd expect if the brain just "registered" inputs
Mismatch Negativity (MMN):
- Play a sequence: beep-beep-beep-BOOP
- The BOOP triggers extra neural activity—a prediction error
- This happens automatically, even if you're not paying attention
Implications for Understanding Subjective Experience
Recognizing the role of expectations has profound implications:
- Individual differences: People with different backgrounds, expertise, or cultural contexts may genuinely experience the same physical stimuli differently
- Plasticity: Subjective experience is not fixed—learning new categories, concepts, or expectations can literally change what we perceive
- Communication challenges: When people disagree about "what they see," they may not be interpreting the same percept differently—they may be having different percepts
- Scientific method: Understanding perception requires controlling not just stimuli but also participants' prior knowledge and expectations
The Pepsi Paradox (shoutout to Lee!)
🎥 Watch: Pepsi Challenge (1982)
🎥 Watch: Pepsi Challenge (2025)
The Marketing Mystery
Researchers and marketers have observed a puzzling phenomenon: in blind taste tests, more people prefer Pepsi, yet Coca-Cola consistently outsells Pepsi in the marketplace. This discrepancy launched the famous "Pepsi Challenge" advertising campaign and even led Coca-Cola to reformulate their product as "New Coke" in 1985—a decision they quickly reversed after public backlash.
McClure et al. (2004)
Neuroscientists used fMRI to investigate what happens in the brain during branded versus unbranded taste tests. Their study revealed distinct neural signatures:
- Blind conditions: When participants didn't know which cola they were drinking, taste preferences correlated with activation in the ventromedial prefrontal cortex (vmPFC), a region involved in reward processing and subjective value.
- Branded conditions: When participants knew they were drinking Coca-Cola, additional brain regions activated, including the dorsolateral prefrontal cortex (dlPFC) and hippocampus—areas associated with memory, cognition, and cultural associations.
- The critical finding: Brand knowledge didn't just add information on top of taste—it changed the subjective experience itself, with different neural patterns of reward processing.
McClure, et al. (2004). Neural correlates of behavioral preference for culturally familiar drinks. Neuron, 44(2), 379-387.
Wine, Price, and Placebo: More Examples of Expectation Effects
Wine Tasting and Price Labels
- This isn't just about reporting bias—neural correlates of subjective pleasure actually changed
- Marketing and pricing information become integrated into the gustatory experience
- Expert wine tasters show similar effects, though sometimes reduced in magnitude
The Placebo Effect
Placebo effects provide perhaps the clearest evidence that expectations produce measurable physiological changes:
- Pain studies: When people believe they've received pain medication (but receive an inert substance), they report reduced pain and show decreased activation in pain-related brain regions including the anterior cingulate cortex and insula.
- Neurochemical changes: Placebo analgesia involves actual release of endogenous opioids, measurable through PET imaging with opioid receptor antagonists.
- Parkinson's disease: Placebo treatments for Parkinson's can trigger dopamine release in the striatum, measurable with neuroimaging.
Branding and Packaging
Additional research has shown that expectations influence experience across many domains:
- Food tastes better when presented on fancy dishware versus paper plates
- Coffee is rated as tasting better from a branded cup versus a plain one
- Luxury brand labels increase reported product satisfaction even when products are identical
- Color of beverages affects reported taste qualities (e.g., red colored drinks taste "sweeter")
Methodological Approaches
How Do We Study Expectation Effects?
Researchers use several complementary methods to isolate expectation from sensory processing:
- Blinded designs: Comparing conditions where participants do and don't have contextual information (brand, price, description)
- Deceptive labeling: Providing false information to examine how expectations diverge from sensory reality
- Neural decoding: Using machine learning to decode what participants are experiencing from brain activity patterns, comparing sensory and expectation-related neural signatures
- Temporal dissociation: Examining when expectation-related brain activity occurs relative to sensory processing
- Computational modeling: Formal models of predictive processing that generate quantitative predictions about neural responses
Challenges and Considerations
- How do we define "veridical" experience when top-down influences are always present?
- Can we ever fully remove expectations, or do we just replace one set of expectations with another?
- Do expert tasters (wine, coffee, etc.) have more "direct" sensory access, or just different expectations?
- What is the relationship between conscious expectation and unconscious prediction?
Discussion Questions
- If our subjective experiences are shaped by expectations, does this mean marketing "manipulates" us, or is it simply providing contextual information that legitimately affects experience?
- What implications does this research have for consumer behavior, medical treatment, and other domains where subjective experience matters?
- Is there a meaningful distinction between a wine that "really tastes good" versus one that tastes good because of the label?
- How might we design studies to determine which aspects of experience are more versus less malleable by expectations?
- What does predictive processing suggest about the possibility of pure, unmediated conscious experience?
Key Takeaways
- Subjective experience integrates bottom-up sensory information with top-down predictions and expectations
- Brand knowledge, pricing information, and contextual cues produce measurable changes in both reported experience and neural activity
- The predictive processing framework suggests perception is inherently inferential rather than direct
- Placebo effects demonstrate that expectations can trigger real physiological and neurochemical changes
- Perception is not passive stimulus processing
- Methodological innovations allow researchers to separate sensory from expectation-related neural responses