Bottom-Up Processing

Imagine you're stumped on what to eat at the food court until a waft of tangy citrus catches your attention. As the scent draws you in, you catch a hint of spicy pepper luring you toward a part of the dining area you haven't perused yet. Is that cilantro? Garlic? Fish sauce? As you search for the source, you notice a Thai food stall serving Spicy Garlic Lime Chicken—yep, that sounds exactly like what you were looking for.

Detecting individual scents, then combining them to identify a dish without context from the bigger picture, is bottom-up processing in action. In this type of attention, we break down our environment's sensory information into its fundamental components, then assemble it to figure out what the scent could be. Notice that in the food court, you didn't use any previous knowledge to arrive at Spicy Garlic Lime Chicken; it was simply the combination of aromas that piqued your interest. This exemplifies how bottom-up processing tends to follow a sort of “what you see is what you get” belief—no learning is required to perceive new stimuli. 

Our perceptions, according to bottom-up processing, are built from direct sensory input without higher processes getting involved. But in reality, once we receive inputs, we also interpret them using prior knowledge, previous experience, and guiding thoughts, which refers to the other side of the coin, top-down processing. For example, smelling a toasted bun and instantly thinking “burgers!” would be exercising top-down processing because past experiences allow you to assume what's letting off that scent by speculating the type of dish that typically accompanies it. In short, bottom-up processing is like building a puzzle piece by piece from the individual elements, while top-down processing is like having the picture on the front of the box to guide you and using that knowledge to piece the puzzle together faster.

Highlighting the key differences between bottom-up and top-down processing can also help to grasp the concept:¹

• Data-driven vs. knowledge-driven processing of information: The brain focuses on the direct data presented by sensory systems in bottom-up processing vs primarily using previously acquired knowledge and context clues in top-down processing. 
• Physical vs. psychological: Bottom-up is linked to the physical stimulation of sensory receptors and transmission of neural signals, whereas top-down taps into cognitive processes like memory retrieval that are much more psychological in nature. 

During bottom-up processing, different physical stimuli—sight, sound, smell, touch, and taste, but also submodalities like temperature, pressure, and proprioception (our ability to sense our position and movement in space)—are detected by our sensory receptors: eyes, ears, nose, skin, and tongue. This input is converted into neural impulses and then sent to the brain and spinal cord for interpretation, a process known as transduction. Different senses have their own respective pathways. As an example, the visual transduction pathway includes the retina, optic nerve, optic chiasm, optic tract, lateral geniculate body, optic radiations, and finally the visual cortex of the brain.²² 

Bottom-up processing is relevant across theoretical explorations of human perception and practical applications in areas such as social cognition, technological innovation, and law enforcement. Over time, research has deepened our understanding of how this perceptual mechanism operates, highlighting both bottom-up's strengths and limitations. As studies have demonstrated the process’s reliability in certain contexts and shortcomings in others, the evolving insights into bottom-up processing have shaped how—and when—we can rely on automatic responses. Since bottom-up processing happens subconsciously without our control, the real choice lies in whether we trust our instincts or decide to seek additional information from top-down processes to confirm them. 

Top-Down Processing: The complementary counterpart to bottom-up processing, where our brains draw upon prior knowledge, experiences, and context in order to interpret incoming sensory information.

Sensory Modalities: The different ways we can receive and process information through our senses, including sight, sound, touch, smell, and taste, but also submodalities like temperature, pressure, and proprioception (our ability to sense our position and movement in space). 

Sensory Input: The response of a sensory receptor (eyes, ears, nose, tongue, skin) when it receives stimuli, like hearing a loud noise, feeling a rough surface, or smelling a sweet scent. 

Stimulus-Response: Simply the interaction between a stimulus and an organism's response to it. Stimulus refers to an external or internal change that can be detected by a receptor, and response is the reaction that we display to that change, like squinting when we look up at the sun.²

Perception: The organization, interpretation, and identification of sensory input in order to create a mental impression and form an understanding of information in our environment. 

Organisms with basic cognitive abilities have engaged in bottom-up processing since the dawn of time, as this fundamental ability to detect environmental stimuli has been essential for survival across all species. However, its formalization as a defined concept is rooted in the early theories of empiricism, initially articulated during the Renaissance by Francis Bacon,³ who argued that knowledge is derived from experience and observation. John Locke continued to develop the empiricist philosophy during the Age of Enlightenment by proclaiming the mind as a “blank slate” (tabula rasa) at birth and that all knowledge is acquired through various sensory experiences. Setting the stage for future research, both Bacon and Locke's emphasis on direct sensory experience as the foundation of knowledge mirrored the principles of bottom-up processing.⁴ 

Later on in the 1870s, psychologist Wilhelm Wundt pioneered the concept of structuralism, an approach that suggests that breaking down phenomena into their different components (from the bottom up) is the best way to analyze the underlying meaning involved in human thought.⁵ By introducing structuralist theory, Wundt captured the essence of bottom-up processing and helped popularize the idea that human thought is built from individual details to a broader understanding. 

Later thinkers took this further during the behaviorism era of the 1920s to 1950s, led by John B. Watson and B.F. Skinner. These psychologists believed that observable behavior, and that only what can be seen and measured, should be studied. Watson emphasized stimulus-response relationships,⁶ implying a bottom-up perspective and the irrelevance of mental states in behavioral outcomes, neglecting the existence of top-down processing entirely. 

The rise of cognitive psychology in the 1950s marked a deliberate shift towards attempting to understand internal processes. In 1959, neurophysiologists Hubel & Wiesel provided biological evidence for bottom-up perception when they discovered neurons in the visual cortex that respond to specific features of stimuli like edges and angles.⁷ Their findings indicated the importance of considering both internal and external operations when thinking about how perception is constructed. 

Eleanor J. Gibson was an American psychologist fascinated by early learning in children. In 1960, she set out to determine whether depth perception in infants is innate or not using a platform with a cliff illusion. When most infants hesitated or refused to crawl over the deep side, it supported her perceptual learning theory,⁸ which suggests that perception arises from the interaction between raw sensory input (bottom-up processing) and prior experience or learning (top-down processing).

Coining the term “bottom-up processing” wasn't attributed to a single person in particular, but it was Eleanor Gibson and her husband James Gibson who popularized the concept through their research, establishing it as a fundamental topic in cognitive psychology.²³ Their findings support the contemporary perspective that widely accepts that both types of attention are essential for perception. Sensory input provides the raw data, while prior knowledge, experience, and context help us interpret and make sense of it. Today, bottom-up processing plays a key role in machine learning and applications of AI, because these systems rely heavily on raw data input. However, because these tools often lack the complex cognitive top-down capabilities that we humans have, their dependence on bottom-up processing alone has highlighted their limitations when used in isolation, and serves as further proof that humans rely on bottom-up and top-down processing working in conjunction.

Wilhelm Wundt 

A mid-19th-century German physiologist and psychologist, acknowledged as the founder of experimental psychology. In line with the principles of bottom-up processing, he emphasized breaking down consciousness into basic sensory elements to form an understanding, establishing perception as a constructive process beginning with raw sensations. 

David Hubel & Torsten Wiesel 

In 1981, these researchers won the Nobel Prize in Physiology or Medicine for their research on feature detection theory. They proposed that the nervous system breaks down complex stimuli, which are then analyzed by specialized neurons called feature detectors. These findings showed that the brain constructs perception by starting with basic features from external stimuli, gradually assembling them into a complete understanding of an object or environment.

Eleanor J. Gibson

An American psychologist who focused on reading development and perceptual learning in infants. At the core of her reading research was the idea that learning begins with sensory input and develops through exposure and active exploration, rather than relying primarily on prior knowledge or top-down expectations. Her 1960 Visual Cliff Experiment9 was evidence for the dynamic interaction between bottom-up and top-down processing in depth perception and perception in general. 

Our understanding of bottom-up processing's rapid, subconscious operations and stimulus-driven responses has, in some ways, enabled us to create a more efficient and responsive society. Many of our real-world systems capitalize on the speedy and all-encompassing characteristics of bottom-up processing to enhance social welfare, marketing strategies, and safety procedures, keeping the world turning. 

Eyewitness testimony

You're walking down the street, and suddenly you see someone snatch a purse and dart into an alley. You're caught off guard, and your heart is racing—but certain sensory details make an imprint instantly.

When witnessing a crime in real-time, we tend to remember distinct features of a suspect, like a green bag they carried or a tattoo on their right arm. In the moment, we don't really understand what's going on—what the motive is—but these raw sensory details tend to stick. 

As witnesses, we are asked to recall a direct timeline of events. Our recollections are primarily based on visual and auditory signals that stood out at the time, since, when witnessing the crime, we didn't have access to prior knowledge at that moment, the sensory input we gathered forms the basis of our timeline. Our lack of contextual understanding results in a dependency on bottom-up processes, causing us to overlook the broader sequence of events, ultimately enhancing our objectivity as a witness. 

However, in practice, using our bottom-up processing to later identify a suspect is not foolproof because identification usually involves more than just sensory recall. This is explained by signal detection theory (SDT)—witnesses are required to distinguish the “signal” (or in this case, the suspect) from the sensory “noise” (like similar-looking individuals in a lineup or hostile voice tones from law enforcement). Our detection skills do, in fact, suffer at the hands of sensory noise. The American Psychological Association estimates that a whopping one in three eyewitnesses makes an erroneous identification, showing that despite accurate initial perception, we are prone to later misjudgment.¹⁰ 

Safety and emergency response

No matter where we are in the world, one of the most universally recognizable sensory cues is the blaring of a siren coupled with flashing overhead lights. Even if we're driving on autopilot—tuned in to the rhythm of the radio or brainstorming ideas for dinner—the second we spot those red and blue flashes out of the corner of our eye and make out the faintest wail of a siren, we automatically react by pulling over without hesitation. No conscious thought is required, and at that moment, our ability to rely on the rapidity and accuracy of bottom-up processing might just save a life!

Beyond traffic situations, we can depend on the automatic exertion of bottom-up processing to prevent accidents in a split second. Studies have shown that our response times average within 200 milliseconds of a sudden sound or visual cue, highlighting the degree to which reactionary stimulus-response relationships are deeply ingrained. The success of an initial reactionary response in hospitals or factories— where alarms, flashing monitors, or sudden equipment malfunctions demand immediate attention—is necessary for enabling secondary, more top-down responses needed to deliver emergency care or address machinery failures effectively.

Attention first, logic later marketing

Marketers design ads that appeal to our five senses before we can even realize it, using bottom-up processing to influence us without our conscious awareness. This opens up questions, like do we find ourselves randomly reaching for a snack in a bright bag or with oversized lettering because we prefer that flavor, or simply due to the flashy packaging instantly grabbing our attention? Or, do we keep buying the same brand of shampoo because the commercial says it keeps our hair healthy, or do we just like the catchy and uplifting background music in the ad? Could be the latter. Marketers love to exploit our bottom-up responses to drive impulse buys—so beware! 

Using a combination of both bottom-up and top-down processing can allow us consumers to stay mindful and less susceptible to marketing tactics. For example, that Coke bottle made its way into our shopping cart because, for some reason, we took notice of the label (bottom-up). Two minutes later, we put it back, realizing it's because it randomly read the name of our middle school soccer coach, whom we haven't seen in 15 years (top-down). We might think we want to buy a box of Lucky Charms because “They're Magically Delicious,” but our top-down logic reminds us that that jingle comes on every 20 minutes during the ad break of our favorite TV show. Stay alert!

Bottom-up processing takes the cake when it comes to precise data collection under controlled conditions. However, banking on this type of attention in complex or ambiguous situations can be risky. Understanding its limitations is crucial to avoid critical errors in perception and decision-making, especially in high-stakes environments.

Overwhelmed by input: sensory overload

During sensory overload, when we are taking in more sensory information than our brain can process at one time, we tend to inadvertently shut off our top-down processing mechanism in an effort to cope. This could lead to confusion and difficulty making decisions, often resulting in choice paralysis, which is when an overwhelming number of choices makes it more difficult to actually make one. 

The unfortunate reality of bottom-up processing is our inability to consciously stop it from happening. For first responders in a burning building, for example, when their senses are being hit from every angle, surrounded by smoke, screams, alarms, flashing lights, and falling debris, they may be more susceptible to split-second mistakes, like missing a safe exit point or misjudging which victim needs help first. Competing stimuli can cause us to make judgment errors without a composed top-down analysis to supplement our decision-making.

It's not always the quantity of sensory information, but sometimes the quality that can contribute to overload. A study published in the Journal of Trauma (2010) found that ER doctors sometimes bypass the standard Airway-Breathing-Circulation (ABC) trauma protocol when confronted with visually dramatic injuries like gory wounds. Physicians were more likely to focus on the obvious problem despite the patient having a more life-threatening, less visible issue like a collapsed lung.¹¹ Responders reacting to what stands out first is, unfortunately, a heuristic driven by bottom-up processing. 

Why bottom-up literal thinking isn't enough

What if we couldn't read the room? When communicating with others, failing to pick up on important emotional cues that aren't immediately obvious can sometimes be detrimental. For instance, Sheldon Cooper provides plenty of cringeworthy moments in The Big Bang Theory when he's unable to gauge the mood. To a clearly scoffing, “Oh great, this is just what I needed today,” he responds with an enthusiastic, “I'm glad to be of service,” acknowledging the surface meaning,¹² but not the actual sarcastic meaning because of his brain's dominant reliance on bottom-up processing. Lucky for Sheldon, the humor and charm in his missteps often make them endearing. However, in real life, ignoring social context can lead to being seen as out of touch or, worse, a social pariah. 

What if we struggle to read between the lines? Effective communication depends on the top-down integration of prior knowledge, expectations, and context to interpret complex indirect speech acts, irony, or sarcasm. It would be impossible to understand George Orwell's famous line in Animal Farm, “All animals are equal, but some animals are more equal than others,” without recognizing the broader context of Soviet communism and the satirical tone as a critique of political hypocrisy and authoritarianism. A powerful and thought-provoking line often only exists within its context; the bottom-up and literal interpretation that “yes, some animals are more equal than others” is so ridiculous that it sounds like bad math, not political commentary. 

AI and machine vision

A self-driving car sounds great and all—until it fails to recognize a stop sign that's covered in graffiti.¹⁶ Numerous AI systems use image recognition algorithms that rely on accurate bottom-up processing, similar to humans. Since racking up expensive tickets and causing car accidents isn't exactly part of most of our dreams for the future, we need to be careful when developing AI that lacks the top-down reasoning that humans use to fill in informational gaps and evaluate ambiguity. 

Many AI systems can perform impressively well in controlled settings but struggle to adapt to the unpredictable real world, where voice tone needs to be considered, cultural implications exist, and interpreting obscure objects can be crucial. It's one thing when your face ID won't unlock your phone when you're wearing sunglasses. It's another when AI-interpreted chest scans for lung cancer are saved without a visual box showing the area of concern, and radiologists missed diagnoses, causing false negatives to rise to 33% and false positives to 86%.¹⁷

The superpower of bottom-up thinking in Autism Spectrum Disorder

Plenty of influential figures throughout history—like Albert Einstein, Emily Dickinson, and Andy Warhol—exhibited characteristics commonly linked to autism and are celebrated for their groundbreaking creativity and unconventional problem-solving. One trait often shared with people on the spectrum is that they exhibit a bottom-up approach to thinking, meaning they process information starting with details first in order to build an overarching understanding or concept.¹⁸ Autistic individuals are often referred to as “specialized thinkers” because their detail-oriented style allows them to break down complex ideas into manageable parts, excel at memorizing large quantities of data, and pick up on nuanced patterns, whether it be in music, language, or numbers. 

This manner of thinking also means that autistic people may be less subject to the framing effect,¹⁴ which occurs when people react differently due to the context in which information is presented; for example, those on the spectrum might focus on what is actually being said instead of the positive or negative way it's framed—instead of a product being 90% fat-free, they might fixate on the fact that the product still contains 10% fat. 

However, for people with or without ASD, bottom-up thinking can present some challenges. Difficulty with generalizations and applying information from one specific instance to another similar one is a common struggle. Social cues can be easily misread when a greater social context is hard to bear in mind, and constant attention to detail can easily lead to sensory overload and trouble filtering out irrelevant information. Nevertheless, the ability to process information so carefully is a powerful asset that often translates into unique insights and innovative thinking. 

Ethnoastronomy

Different cultures perceive different constellations from the same star patterns because creating a constellation is a bottom-up process. The study of how constellations have evolved into metaphorical representations of animals, mythological creatures, or inanimate objects is called ethnoastronomy. Naming constellations helped explorers and navigators identify landmarks in the sky, keep up with calendars to track agricultural seasons, and played a significant role in cultural ceremonies, celebrations, and storytelling.¹⁹

Polynesians use stars for navigation by observing the patterns of the stars, their rising and setting points, and changes in light intensity. Navigators then build a mental star compass from raw visual input: what they see and experience night after night, learning which stars appear during which seasons by accumulating sensory observations over time. These observations can be pieced together to create navigation tools like the Hawaiian star compass, developed by master navigator Nainoa Thompson, which has been used to direct the 2500-mile voyage from Hawaii to Tahiti without modern technology several times.²⁰

Aboriginal Australians use the dust lane shapes of the Milky Way to signal when emus are laying their eggs. Through pattern recognition of raw visual features, they noticed the shape of a giant emu formed by the dark gaps between the stars in the Milky Way. Initially, no cultural template was imposed—this emu shape emerged from bottom-up processing of the natural night sky. Later, top-down processing comes into play when they learn that in April, the Emu in the Sky takes on a certain orientation, which coincides with when real emu mothers lay their eggs. This constellation is now deeply tied to Aboriginal culture and serves as the focal point for seasonal knowledge and ceremonial practices.²¹

Purely through seeing, noticing, and connecting, human perception and attention to detail can give rise to complex astronomical systems—all from the grounds of direct sensory output. The fact that different cultures at different points in time draw varying uses from the same sky patterns highlights how bottom-up processing can lead to unique symbolic creations and imaginative expressions.

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

By recounting a vivid childhood memory, our writer Preeti Kotamarthi reveals how our minds navigate attention and choose what we want to focus on. Ultimately, she emphasizes why true balance between top-down and bottom-up processing is key to creating experiences that feel seamless rather than overwhelming, whilst also outlining what different scientific perspectives have to say about the bottom-up vs. top-down discussion. 

Top-Down Processing

Discover the counterpart of bottom-up processing—top-down processing—and how the two work together to help us navigate everyday life. This piece explains how our expectations, experiences, and focus shape what we perceive, as well as explores real-world applications that reveal the power of this dynamic duo. 

1. Connor, C. E., Egeth, H. E., & Yantis, S. (2004). Visual attention: Bottom-Up versus top-down. Current Biology, 14(19). https://doi.org/10.1016/j.cub.2004.09.041 
2. The Editors of Encyclopaedia Britannica (2023, April 12). stimulus-response theory. Encyclopedia Britannica. https://www.britannica.com/science/stimulus-response-theory
3. Klein, J. (2003). Francis Bacon (Stanford Encyclopedia of Philosophy). [online] Stanford.edu. Available at: https://plato.stanford.edu/entries/francis-bacon/.
4. Uzgalis, W. (2022, July 7). John Locke. Stanford Encyclopedia of Philosophy; Metaphysics Research Lab, Stanford University. https://plato.stanford.edu/entries/locke/
5. Kim, A. (2006, June 16). Wilhelm Maximilian Wundt (Stanford Encyclopedia of Philosophy). Stanford.edu. https://plato.stanford.edu/entries/wilhelm-wundt/
6. Loo, J. (2021). Behaviorism. The Decision Lab. https://thedecisionlab.com/reference-guide/neuroscience/behaviorism
7. Lienhard, D. (2017). David H. Hubel and Torsten N. Wiesel’s Research on Optical Development in Kittens | The Embryo Project Encyclopedia. Asu.edu. https://embryo.asu.edu/pages/david-h-hubel-and-torsten-n-wiesels-research-optical-development-kittens
8. Adolph, K. E., & Kretch, K. S. (2015). Gibson’s Theory of Perceptual Learning. International Encyclopedia of the Social & Behavioral Sciences, 10(2), 127–134. https://doi.org/10.1016/b978-0-08-097086-8.23096-1
9. Rodkey, E. (2011, August). The woman behind the visual cliff. Https://Www.apa.org. https://www.apa.org/monitor/2011/07-08/gibson
10. Sagana, A., Sauerland, M., & Merckelbach, H. (2013). Witnesses’ Blindness for their Own Facial Recognition Decisions: A Field Study. Behavioral Sciences & the Law, 31(5), 624–636. https://doi.org/10.1002/bsl.2082
11. Ferrada, P. (2023). Prioritizing Circulation to Improve Outcomes for Patients with Exsanguinating Injury: A Literature Review and Techniques to Help Clinicians Achieve Bleeding Control. Journal of the American College of Surgeons, 238(1), 129–136. https://doi.org/10.1097/xcs.0000000000000889
12. Hopkins, M. (2021, September 6). The 1%: Autism in The Big Bang Theory. The Oxford Blue. https://theoxfordblue.co.uk/the-1-autism-in-the-big-bang-theory/
13. Pilat, D., & Krastev, S. (2021). Framing effect - Biases & Heuristics. The Decision Lab. https://thedecisionlab.com/biases/framing-effect
14. Simons, D. J., & Chabris, C. F. (1999). Gorillas in Our Midst: Sustained Inattentional Blindness for Dynamic Events. Perception, 28(9), 1059–1074. https://doi.org/10.1068/p281059
15. Picasso, P. and Ashton, D. (1972). Picasso on Art.
16. Matthew Sparkes (2022). Tesla recalls 50,000 cars that disobey stop signs in self-driving mode. [online] New Scientist. Available at: https://www.newscientist.com/article/2307147-tesla-recalls-50000-cars-that-disobey-stop-signs-in-self-driving-mode/.
17. Bernstein, M. (2023). Can incorrect artificial intelligence (AI) results impact radiologists, and if so, what can we do about it? A multi-reader pilot study of lung cancer detection with chest radiography. Can incorrect artificial intelligence (AI) results impact radiologists, and if so, what can we do about it? A multi-reader pilot study of lung cancer detection with chest radiography, [online] 33(11). doi:https://doi.org/10.1007/s00330-023-09747-1.
18. Texas Health & Human Services Commission (2024). Module 4: Common Behavior Characteristics of Individuals with Autism Spectrum Disorder (ASD) | Texas Health and Human Services. [online] www.hhs.texas.gov. Available at: https://www.hhs.texas.gov/services/disability/autism/autism-training-opportunities/autism-spectrum-disorders-training-program-first-responders/module-4-common-behavior-characteristics-individuals-autism-spectrum-disorder-asd.
19. Gullberg, S.R. (2019). Cultural Astronomy for Inspiration. Proceedings of the International Astronomical Union, 15(S367), pp.265–268. doi:https://doi.org/10.1017/s1743921321000612.
20. Polynesian Voyaging Society (2023). Polynesian Wayfinding. [online] Hōkūleʻa. Available at: https://hokulea.com/polynesian-wayfinding/.
21. University of Kent. (2017). Emus in space: visualising western and indigenous knowledge. [online] Science Comma. Available at: https://blogs.kent.ac.uk/sciencecomma/2017/05/17/emus-in-space/.
22. SciencDirect (2012). Visual Pathway - an overview | ScienceDirect Topics. [online] Sciencedirect.com. Available at: https://www.sciencedirect.com/topics/neuroscience/visual-pathway.
23. Psychology Town (2024). Understanding Perceptual Development: Theories and Insights • Psychology Town. [online] Psychology Town. Available at: https://psychology.town/general/perceptual-development-theories-insights/ [Accessed 27 May 2025].