Top-Down Processing

Have you ever proofread something over and over, just to glance at it later and see an obvious typo staring back at you? Or maybe you’ve been on the receiving end of a poorly spelled message that you’re surprisingly still able to understand. These examples show that reading is more than recognizing letters and words—it’s about applying our existing knowledge to understand meaning.

Top-down processing is the mechanism that allows us to subconsciously swap letters within a word or tune out background noise when we’re focused. Unlike bottom-up processing, which relies on the stimuli around us to build our mental representations of the world, top-down processing recruits our memories, goals, and knowledge to shape our experiences.¹ The effect is at the level of perception; in the case of sneaky typos, our expectations influence our conscious experience, causing us to miss those minor errors. 

Top-down processing gets its name from the literal and figurative model of how information is processed in the human brain. Information from our senses typically enters the brain from below, near the spinal cord. From there, it travels up and forward through the brain, undergoing progressively more complex processing, until it reaches the frontal areas that are responsible for higher-order functions such as decision-making, impulse control, and motivation.^1,2 The “top” in “top-down” refers to these higher-order functions.

But, perception isn’t a one-way street. Our frontal brain areas are always in communication with the lower-order systems (lower in the anatomical sense and the conceptual sense) to help them work more practically and efficiently.  Not every task needs our full, undivided attention or a microscopic level of precision. If we’ve seen the same word thousands of times before, our brains can often assume that a missing letter doesn’t completely change its meaning, especially using context clues from the rest of the sentence. That way, we can efficiently grasp the big idea rather than getting caught up in the details. These top-down signals are a key contributor to the delicate balancing game between speed and accuracy that is constantly playing out in our brains. 

Bottom-Up Processing: A perceptual process starting from the basic features of sensory stimulus. For example, viewing a novel image by first looking at edges, lines, and color contrast. Bottom-up processes complement top-down processing by working in reverse; using stimuli to update our existing knowledge. 

Sensation: The physical interactions between sensory receptors in our bodies and environmental stimuli. While many senses, like vision, arise from external stimuli, others respond to signals inside our bodies, such as our sense of balance.4

Perception: The process of selecting, organizing, and interpreting sensations.4 If sensation allows you to detect flavors on your tongue, perception describes your ability to recognize and experience the taste of your favorite (or least favorite) foods.

Limited Capacity: The idea that our brains have limited resources for processing information and making decisions.5 Cognitive mechanisms like top-down processing help us conserve valuable processing time and capacity by filling in perceptual gaps and directing us toward our goals. 

Schema: A mental framework constructed from prior knowledge that guides perception and decision-making. Schemas are a form of top-down processing where memories help structure our understanding of the world. 

Constructivist Theory of Perception: A top-down theory of perceptual processing that describes how perception is actively constructed by integrating sensory information, past experiences, and other higher-order cognitive functions..6

Direct Theory of Perception: A bottom-up theory of perceptual processing that rejects the involvement of prior knowledge in perception. Direct theories suggest that our perception relies only on sensory information.

Gestalt Principles: Perceptual rules that we use to organize visual information into meaningful wholes and groups rather than collections of smaller components. Gestalt principles help us simplify complex images using top-down processing to guide perception.

In some ways, the study of perception is older than psychology itself. Before psychology emerged as a distinct discipline in the late 1800s, perception was a central topic of philosophical inquiry for thinkers in ancient civilizations across the world. Philosophers such as Plato and Aristotle extensively questioned whether we should trust our senses to show us an accurate representation of the “real” world (if such a thing exists at all). Plato’s Theory of Forms famously outlines a fundamental idea within modern behavioral science: we can’t always trust our senses.

The philosophical foundation of top-down processing can be traced from ancient societies to the Enlightenment. In 1781, German philosopher Immanuel Kant published Critique of Pure Reason, which argued that the mind plays an active role in shaping our perceptions. Concepts like space, time, and cause and effect are pre-existing constructs—what he called “a priori knowledge”— that shape how we interpret information from the senses.⁷ 

Early philosophical questions about our mind and senses formed the beginnings of psychology as we would recognize it today. But before psychology became its own field, physiologists and philosophers began studying vision to learn more about perception. Between revolutionary contributions to mathematics, meteorology, and physics (ever heard of the conservation of energy?), “post-Renaissance Renaissance Man” Hermann von Helmholtz developed a foundational theory of perception from his study of optics.⁸ von Helmholtz coined the famous phrase “perception is unconscious inference,” emphasizing the importance of prior knowledge and informed guesswork in perception. 

Immanuel Kant

One of the most pivotal figures of modern Western philosophy. During the Enlightenment (late 17th century to early 19th century), Kant published several foundational works on morality, consciousness, and perception.⁷ He was an early proponent of active perceptual processing, theorizing that the mind uses pre-existing constructs to structure perception and laying the groundwork for later theories of top-down processing.

Hermann von Helmholtz

A German physiologist, physicist, mathematician, and philosopher, Hermann von Helmholtz transformed the field of vision science with his invention of the ophthalmoscope, allowing scientists to see the inside of the human eye for the first time.8 Among his many notable scientific contributions, von Helmholtz famously studied the active role of the brain in interpreting visual input.

Richard Gregory

A British psychologist and inventor who originated the concept of top-down processing by arguing that perception is constructive; we actively build our mental representations of the world and integrate past experiences in the process. Inspired by Hermann von Helmholtz’s work on visual perception, Gregory studied optical illusions and the anatomy of vision to develop his constructivist theory of perception.¹⁰

Top-down processes subconsciously shape our understanding of the world around us, affecting almost every aspect of our cognition and decision-making behind the scenes. They help us evaluate options, self-regulate, and work with incomplete information. 

Goal-directed decision-making 

With the abundance of sensory information entering our brains every minute, we need to be able to pick out the material that will help us achieve our goals. This is especially important when we have to evaluate between several possible options. Choice overload explains that the more possibilities, the harder it is to make a decision. Luckily, top-down processes help us weed out irrelevant information by perceptually focusing on stimuli relevant to our motivations and knowing how to efficiently achieve those goals. 

We can see the influence of top-down processing on choice in consumer behavior. Purchasing decisions rely heavily on the buyer’s goals and circumstances—what we buy is informed by knowledge of what we do and do not already have. Research shows that bottom-up processes prevail when evaluating between products of the same type (e.g., different brands of bicycles), while top-down processes lead when the choice is between different products that service the same goal (e.g., deciding between a bicycle, treadmill, or gym membership for exercise).¹² 

Imagine you’re in the market for a new sweater. You’ll probably focus on features like size, texture, and color to decide between brands and styles. However, if you want to stay warm and are choosing between a sweater, a space heater, and a blanket, you might consider factors such as where you often feel cold and the climate of the city you live in, which all depend on memory and reasoning. 

Regulating emotional responses

Our brains are wired to respond rapidly to stimuli that signal danger, but what happens when it misfires and there’s no actual danger? Top-down processing adapts our fight-or-flight responses to scale our emotional responses appropriately, helping us remember that we are safe. For example, imagine you’re watching fireworks on New Year’s Eve. The first “bang!” is loud and sudden, making you jump. But your frontal brain areas tell the sensory systems that there is no real threat, making the following noises subjectively less loud.

Many mental health conditions such as anxiety disorders (social anxiety, PTSD, OCD) and depression are associated with negative emotional responses that are disproportionate to a stimulus’s true threat.¹³ Psychologists and mental health professionals leverage conscious forms of top-down processing through strategies such as cognitive behavioral therapy (CBT), which involves actively restructuring distorted responses by talking and reasoning through them. 

Say, for instance, you’re feeling anxious about an upcoming presentation at work. Even though you know your subject well, you’re nitpicking your presentation and spiralling about the worst possible outcome rather than preparing productively. A cognitive behavioral therapist might talk you through realistic outcomes and how you would overcome them, recruiting higher-order reasoning to scale your emotional response accordingly. As a result, you’re less sensitive to tiny details and better able to focus on the preparation that matters most. 

Resolving ambiguity

Even though we mostly experience the world cohesively, the sensory information we receive is rarely complete. Did you know there’s a blind spot in each of our eyes? You probably wouldn’t just from your perceptual experience; because of the brain’s sophisticated top-down processing, we can best predict what’s there and “fill in the gap.”² Similarly, we can still recognize an object if part of it is obscured; prior knowledge and expectations lead us to perceive it as a whole. Our assumptions tend to follow rules, known as Gestalt principles, to help us conceptualize and group visual features.

Scientists are now using knowledge of top-down processing in the human brain in non-human applications as well. Predictive text and autocomplete functions analyze existing language data, much like we use our knowledge of language patterns to predict the rest of our words and sentences before they’re finished.¹⁴ Machine learning models are trained to detect and recognize certain images using huge datasets, mimicking the wealth of experience and knowledge a person accumulates throughout their lifetime.¹⁵ While we haven’t yet figured out how to measure or recreate the conscious experience of perception, it’s clear that top-down processing plays a significant role. 

While top-down processing is essential for helping us learn and conserve mental resources, we don’t always get things right. Errors in this system can range from inconsequential to costly when assumptions go unnoticed or unchecked. 

Mental rigidity

Relying too heavily on past experiences can make updating assumptions or integrating new information difficult. Top-down processing can contribute to functional fixedness, making it difficult to override expectations with new interpretations of what a stimulus is or can do. This can pose limitations on problem-solving, creativity, and flexible thinking, especially in familiar contexts where prior experience vastly outnumbers new sensory information. 

The expression “if all you have is a hammer, everything looks like a nail” comes from functional fixedness; if someone has hammered nails hundreds of times, they might not realize a hammer can be used as a weight or a handle. Top-down influences from past experiences using a hammer and expectations about the hammer’s function prevent the user from seeing it in a new light.

Development and co-existence

Gregory’s theory of top-down processing does not come without criticism. Psychologist James J. Gibson opposed constructivist theories of perception, arguing that perception is sensation—an entirely bottom-up process.¹⁶ His direct theory of perception suggests that we are born with perceptual abilities and don’t need knowledge to make sense of our environment. Gibson and other critics believed constructivists underestimate the amount of sensory information available in our environments, proposing that accompanying sensory information supports perception rather than past experience. Indeed, Gregory’s theory cannot explain how newborns demonstrate several seemingly innate perceptual abilities without prior experience.¹⁷

Today, you’ll be hard-pressed to find evidence disproving the existence of top-down processing altogether; it’s widely understood that top-down and bottom-up processes work together to form perception. However, researchers still disagree on how the two mechanisms co-exist: what order they take, which takes precedence when, how they develop, and so on. 

Implicit biases and harmful expectations

As we’ve seen, our perceptions are not always accurate or representative of experiences outside of our own. Top-down processing can reinforce harmful inequalities through confirmation bias, where our expectations make us figuratively and incapable of seeing the unexpected. Consequently, our assumptions aren’t updated with new information and a dangerous cycle forms. 

This can have life-altering consequences in law enforcement and judicial systems. Witness testimonies can reinforce racial, gendered, and socioeconomic biases if their perception of a crime is influenced by expectations rather than details of the actual event. The effect isn’t just restricted to the perception of the crime either; research shows that individuals biased with emotional stories of crime were more likely to determine that two ambiguous fingerprints matched.¹⁸ While novel fingerprinting technology removes subjective judgment from the equation, the results still demonstrate our susceptibility to background emotions and expectations. 

Perceptual expertise in radiological scanning

Expertise in many perceptual tasks depends heavily on top-down processing—how else would we apply our learning? One of the most researched examples of top-down perceptual expertise deals with radiologists viewing medical scans.¹⁹

While researchers have recreated and tweaked the general study numerous times, the basic design goes something like this: expert radiologists and novices both view the same scans while researchers track measures like their eye movements and how many abnormalities they successfully identify. This way, any differences researchers find can be reasonably linked to the differing amounts of expertise and experience. 

Different iterations indicate that expert radiologists use more efficient search patterns, are more accurate in detecting abnormalities, and show increased sensitivity to medically important abnormalities. Evidently, their medical training and experience guide change how they view new scans top-down, optimizing the efficiency and relevance of visual perception. For radiologists and medical professionals, knowing what to look for can make all the difference. 

Seeing ghosts… in toast?

If you spend enough time on the internet, you’ll probably come across a viral photo claiming to show a face in toast (or some other food item). Jesus, Michael Jackson, Albert Einstein—if you can name it, someone has probably seen it. 

The phenomenon of seeing faces in random patterns or inanimate objects, called face pareidolia, is caused by top-down processing. The ghostly figures result from the brain’s expectations as it assigns meaning to a visual stimulus. In a laboratory setting, experiments show that people are more likely to see faces in randomly generated images when they were told beforehand that some images would contain a face.²⁰ Using brain imaging techniques, scientists also saw that the frontal brain areas were more active when people reported seeing faces. 

The same concept applies to spotting figures in the clouds or animal-shaped potato chips; top-down processing can make silhouettes seemingly appear from thin air. Face pareidolia and other visual illusions are signals of normal top-down processes in the brain, so you can rest easy the next time you see your breakfast staring back at you.

Blue Cars, Robots, and How Your Brain Chooses to See 

What we consciously perceive is closely linked to attention. Like perception, attention can also be driven by our goals and experiences, which has consequences on the choices we make. In this article, Preeti Kotamarthi discusses how top-down/bottom-up processing models can inform digital design and user behaviour.

Perceptual Set

It’s challenging to recall a single memory without at least a few others popping up. Top-down processes integrate many expectations, memories, and goals to inform perception and behaviour. In this article, our writer Emilie Rose Jones breaks down these psychological “sets” and how they powerfully leverage top-down processing. 

1. Gilbert, C. D., & Sigman, M. (2007). Brain states: top-down influences in sensory processing. Neuron, 54(5), 677–696.
2. Gilbert, C. D., & Li, W. (2013). Top-down influences on visual processing. Nature Reviews Neuroscience, 14(5), 350–363. https://doi.org/10.1038/nrn3476 
3. Davies, R. (1951). Tempest-tost. Rinehart. http://catalog.hathitrust.org/api/volumes/oclc/1595669.html 
4. Spielman, R. M., Jenkins, W. J., & Lovett, M. D. (2020). Sensation versus perception - psychology 2e. OpenStax. https://openstax.org/books/psychology-2e/pages/5-1-sensation-versus-perception 
5. American Psychological Association. (2018). Central limited capacity. APA Dictionary of Psychology. https://dictionary.apa.org/central-limited-capacity
6. Gregory, R. (1970). The intelligent eye. London: Weidenfeld and Nicolson.
7. Brook, A., & Wuerth, J. (2020, October 8). Kant’s view of the mind and consciousness of self. Stanford Encyclopedia of Philosophy. https://plato.stanford.edu/entries/kant-mind/
8. Wade, N. J. (2021). The vision of Helmholtz. Journal of the History of the Neurosciences, 30(4), 405–424. https://doi.org/10.1080/0964704X.2021.1904182
9. von Helmholtz, H. (1924). Helmholtz’s treatise on physiological optics. Optical Society of America.
10. Land, M. F., & Heard, P. (2018a). Richard Langton Gregory. 24 July 1923—17 May 2010. Biographical Memoirs of Fellows of the Royal Society, 64, 163–182. https://doi.org/10.1098/rsbm.2017.0034
11. Kinchla, R. A., & Wolfe, J. M. (1979). The order of visual processing: "Top-down," "bottom-up", or "middle-out". Perception & psychophysics, 25(3), 225–231. https://doi.org/10.3758/bf03202991
12. Park, C. W., & Smith, D. C. (1989). Product-level choice: A top-down or bottom-up process? Journal of Consumer Research, 16(3), 289–299.
13. Cleveland Clinic. (2022). Cognitive behavioral therapy (CBT): What it is & techniques. https://my.clevelandclinic.org/health/treatments/21208-cognitive-behavioral-therapy-cbt
14. How to use auto-correction and predictive text on your iPhone, iPad, or iPod Touch. Apple Support. (2025). https://support.apple.com/en-us/104995#:~:text=Use%20predictive%20text&text=As%20you%20type%2C%20you%20can,or%20the%20globe%20icon
15. Awan, A. A. (2023, May 12). What is machine perception?. DataCamp. https://www.datacamp.com/blog/what-is-machine-perception
16. Gibson, J. J. (1972). A theory of direct visual perception. In J. Royce, W. Rozenboom (Eds.). The Psychology of Knowing. New York: Gordon & Breach.
17. Bremner, J. G. (2011). Four themes from 20 years of research on infant perception and cognition. Infant and Child Development, 20(2), 137–147. https://doi.org/10.1002/icd.723
18. Dror, I. E., Péron, A. E., Hind, S.-L., & Charlton, D. (2005). When emotions get the better of us: the effect of contextual top-down processing on matching fingerprints. Applied Cognitive Psychology, 19(6), 799–809. https://doi.org/10.1002/acp.1130
19. Waite, S., Grigorian, A., Alexander, R. G., Macknik, S. L., Carrasco, M., Heeger, D. J., & Martinez-Conde, S. (2019). Analysis of perceptual expertise in radiology - current knowledge and a new perspective. Frontiers in Human Neuroscience, 13, 213. https://doi.org/10.3389/fnhum.2019.00213
20. Liu, J., Li, J., Feng, L., Li, L., Tian, J., & Lee, K. (2014). Seeing Jesus in toast: Neural and behavioral correlates of face pareidolia. Cortex, 53, 60–77. https://doi.org/10.1016/j.cortex.2014.01.013