Dual code: studying with image and text

You highlight the paragraph about the optic nerve. Next to it, a diagram of the eye. On the exam, the sentence comes back; the drawing does not. Or the opposite: you recognize the figure and cannot explain the mechanism. Dual encoding theory starts from a simple idea: the mind handles information better when the verbal channel and the nonverbal channel work together, without tripping over each other. In medical education, this becomes a practical question: how to combine text and image to learn anatomy, physiology and clinical decision-making without overloading working memory (Mayer, 2010).

NEXO is a daily medical diagnosis game: one clinical case a day, six progressive clues, and ICD-10 proximity feedback across five levels.

This text translates the theory for the studying of the student and the resident, links it to the multimedia design principles tested in medical classrooms and shows what usually fails (crowded slides, captions that compete with speech, decorative figures). It is not a post about learning styles: the angle is combining channels with evidence, not labeling the student as visual or auditory. Anyone who has already read active learning and knowledge retention finds here the zoom on the image-text pairing.

Key points

• Dual encoding describes verbal and nonverbal representations that reinforce each other when the instructional drawing respects the limits of working memory (Mayer, 2010).
• In medical classes, applying multimedia principles (coherence, signaling, contiguity, modality) improved retention and transfer in studies by Issa and colleagues (Issa et al., 2011; Issa et al., 2013).
• An image without a verbal task (explaining, labeling, predicting) becomes passive recognition; text without a figure leaves what is spatial abstract.
• Active retrieval remains essential: dual encoding does not replace the testing effect (Larsen et al., 2013; Raupach et al., 2016; Dunlosky et al., 2013).
• Poorly designed cognitive load (decoration, redundancy, everything at once) takes resources away from what should build the mental model (van Merriënboer and Sweller, 2010).

What dual encoding is

Allan Paivio described the idea that the mind uses systems for words and for mental images, and that learning improves when the two connect. In the line applied to instruction, Mayer translates this into the design of classes and materials: words and figures should form a coherent mental model, not two parallel currents that compete (Mayer, 2010).

It is not I am visual

Saying I learn better with images as a fixed personality trait is another debate (the myth of styles). Dual encoding says something else: for almost everyone, combining the verbal and nonverbal channels well helps, and combining them badly gets in the way. The focus moves away from the student's label and toward the design of the task.

Verbal and nonverbal in medicine

• Verbal: definition, mechanism, diagnostic criterion, management script.
• Nonverbal: anatomy, the path of a catheter, pressure curve, ECG pattern, map of hypotheses.
• Mixed: a clinical case with a textual clue and an imaging finding that must be described in words.

Clinical practice mixes the two all the time. That is why studying only text or only figures leaves one half weak. Differential diagnosis with clues already forces you to join finding and label; here the point is to design the material for that junction.

Link to cognitive load

van Merriënboer and Sweller apply cognitive load theory to education in the health professions: the design should reduce extraneous load (noise in the material), manage intrinsic load (difficulty of the content) and favor germane load (schema building) (van Merriënboer and Sweller, 2010). Well-executed dual encoding is a way to favor germane load without blowing up working memory.

Working memory and overload

Working memory holds only a few items at a time. If the slide shows a dense diagram, a block of text identical to what the teacher says, and three decorative animations, the channels get congested. Mayer calls attention to reducing what the design adds without need, so that resources remain for the content and for building the model (Mayer, 2010).

Practical implication

Before adding a figure, ask: does it reduce spatial ambiguity or does it just decorate? Before pasting the whole paragraph onto the slide, ask: will the speech and the text compete in the same verbal channel?

Spatial and temporal contiguity

A word and its corresponding figure should appear together. Caption far away, figure on one page and text three screens later: the student spends energy hunting for the connection, not building the model. In your notebook, draw the arrow and write the name on the same line.

Segmenting

Breaking the mechanism into steps (first the part, then the flow, then the clinical failure) reduces the peak load. Useful in metabolic pathways, conduction circuits and management algorithms. The same logic appears when you study a protocol: one step at a time, with the flow figure next to the short sentence of the criterion.

In the practice of a study shift, if the feeling is one of fog after ten minutes of slides, the cause is usually extraneous overload, not a lack of intelligence. Stop, remove one layer (text or figure), and resume. This diagnosis of your own material is metacognition applied to method (metacognition in medicine).

Multimedia principles in medicine

Mayer summarizes principles that Issa and colleagues brought into the medical and surgical classroom (Mayer, 2010; Issa et al., 2011; Issa et al., 2013). Yue and collaborators analyzed medical animations under the same theory and showed where design fails in practice (Yue et al., 2013).

Coherence

Remove what does not teach. Landscape photo, generic clipart, giant logo: all of it is excess. An object of the topic (anatomical section, flow, decision diagram) stays. The NEXO blog standard on body photos follows the same logic: a concrete object of the topic, never an empty metaphor.

Signaling

Highlight what matters: an arrow on the nerve branch, color only on the current step of the algorithm. Too much signal becomes noise.

Redundancy

Avoid reading the paragraph identical to the projected text. Either the speech adds something, or the text is trimmed. Two identical versions compete in the verbal channel.

Modality

In many contexts, narrating while the figure appears works better than piling written text on top of the same figure. There are limits: new terms and dense lists still need a written anchor.

Pre-training

Teach the names of the parts before the full circuit. Without vocabulary, the pretty figure does not become a mental model.

Principle	What to do in the notebook or on the slide	What to avoid
Coherence	Only figures that teach the core point	Decoration and empty metaphor
Contiguity	Caption attached to the part	Text far from the image
Signaling	One arrow at the critical point	Five colors with no hierarchy
Redundancy	Speech or text, not both identical	Reading the slide word by word
Segmenting	Step by step through the mechanism	Everything at once on the same screen
Pre-training	Names of the structures first	Full circuit without vocabulary

What classroom evidence showed

Issa and colleagues redesigned classes with multimedia principles and measured learning in medical students. In 2011, applying the principles improved performance compared with the traditional format at the same institution (Issa et al., 2011). In 2013, the group examined comprehension with multimedia design and reinforced that the gain is not cosmetic: it changes retention and transfer when the material respects working memory (Issa et al., 2013).

Mayer argues that the science of learning should inform medical education with the same rigor that clinical practice demands for treatments (Mayer, 2010). Cook and colleagues, in reviews on internet-based learning in the health professions, show that instructional design (not just the digital medium) explains much of the variation in effect (Cook et al., 2010). Ruiz and colleagues discuss the impact of e-learning on medical education with the same care: a new medium does not guarantee a gain if the design is poor (Ruiz et al., 2006).

Honest limits

Not every figure helps. A decorative figure can make things worse. Not all content is spatial: ethics, communication and some lab criteria call for more verbal material and examples. Well-designed multimedia without active retrieval still fades: Larsen, Raupach and Dunlosky reinforce testing techniques and retrieval practice (Larsen et al., 2013; Raupach et al., 2016; Dunlosky et al., 2013).

Medical animations deserve a critical look: Yue et al. show that many violate basic principles (too much information, weak signaling, inadequate pace) (Yue et al., 2013). Before adopting a long video, check whether it segments, signals and avoids redundancy.

In residency, the same principle applies to the lecture slide deck: if the class only photographs the slide and cannot redraw the circuit on the next shift, the material trained recognition, not transfer. Ask for a blank sketch at the start of the next class. The cost is low; the diagnostic value of the method is high.

How to study with image and text

Anatomy and procedures

1. Look at the figure in silence and anticipate the names.
2. Cover the labels and speak out loud.
3. Redraw the circuit from memory in thirty seconds.
4. Only then check the atlas.
5. Write one sentence of mechanism that links the pieces.
6. The next day, retrieve without looking and mark what failed.

Physiology and curves

Draw the curve before reading the explanation. Then read. Then explain the curve to a colleague without looking. The sequence forces the nonverbal channel to work before passive recognition. If you study alone, record a one-minute audio explaining the curve and listen to it on your commute; then try to redraw it without the audio.

ECG, imaging and skin

Describe the pattern in words before jumping to the label. A label without a description becomes bingo; a description without a label stays incomplete. The two channels close the loop. For skin, force location, color, border and associated symptom out loud before the name of the lesion.

Clinical cases

Build a mini map: on the left, clues (text); on the right, a mental image of what you expect on physical exam or imaging. Update both with each new clue. This connects with case-based learning and with virtual patients.

Flashcards

• Front with a partial figure; back with the name and one line of function.
• Front with a definition; back with a sketch.
• Front with a clinical finding; back with what you expect to see on the image.
• Avoid cards with only long text if the content is spatial.
• Mix in spaced retrieval: do not review the whole deck every day in the same order.

A one-hour routine

1. Fifteen minutes: pre-training names with the figure covered.
2. Twenty minutes: mechanism in segments with your own drawing.
3. Fifteen minutes: short cases linking clue to mental image.
4. Ten minutes: retrieval test without looking at the material.

Common material mistakes

• A slide with a full paragraph and a tiny figure in the corner.
• A video with captions identical to the narration plus an animated banner.
• A colorful atlas you never cover the label on.
• A visual summary that is just a screenshot of an illegible table.
• An infographic with ten typefaces and no hierarchy.
• A figure from another context just because it looks nice (violates coherence).

Quick checklist

1. Does the figure teach a spatial or causal node?
2. Are text and figure together?
3. Is there minimal, clear signaling?
4. Am I reading the same text that is already on the screen?
5. Did I segment the mechanism?
6. Will I test from memory afterward, not just reread?
7. Did I remove the decorations?

Connection with educational games

Games and simulators mix case text and visual representation. Good design applies the same principles: clue and image of the finding together; no noise; with feedback that explains. See also games for medical students, gamification in medical education and ethics in game design.

Dual encoding is no excuse not to write

Explaining in writing what the figure shows closes the verbal loop. Without that, only recognition remains. Ask yourself for a five-line paragraph after each key figure. The error notebook can hold the sketch that failed during retrieval.

For those who teach

• Ask for the drawing from memory before the illustrated answer key.
• In image rounds, force the verbal description before the report.
• On slides, cut text you are going to say.
• Test the material: if the student can only recognize and not explain, the verbal channel is weak.
• Align feedback with the process (feedback on cases), not just with the grade.
• In leagues and groups, alternate who draws and who narrates each case.

What changes in one week of habit

If you change only one thing, change this: never review a figure without covering the caption on the first pass. On the second, write the mechanism sentence. On the third, recall it without the book. In seven days, the notebook shows where the verbal or nonverbal channel still fails. That local diagnosis is worth more than switching atlases.

Teaching example (fictional)

Fictional case with a real clinical presentation in the study of mediastinal anatomy: the student highlights the textbook in yellow, watches a five-minute video with captions identical to the narration and declares that they already know it. The next day, they cannot draw the relationship between the aortic arch and the recurrent laryngeal nerve. The problem was not a lack of hours: it was the verbal and nonverbal channels competing in the video, with no redrawing task and no retrieval. The remedy aligned with dual encoding is simple: pause the video at each segment, cover the screen, draw, name it out loud and only then continue.

Another example: on the ECG, the student recognizes the pattern because they have seen dozens of screenshots, but freezes when describing what they see before the label. Ask for the description (rate, axis, interval, morphology) in four short sentences; only then the name of the condition. This closes the verbal loop over the visual finding and prepares clinical decisions better than image bingo.

Link to confidence calibration

Recognizing a figure creates quick confidence. Explaining and redrawing calibrates it better. Those who only recognize tend to overestimate what they know, a theme close to confidence calibration and knowledge retention. Use the rule: if I cannot draw it or explain it in a minute, I do not yet know enough for the exam or for the shift.

What to ask the preceptor

• One figure per key concept, not a wall of them.
• Time to redraw before the answer key.
• A transfer question: the same structure in another section or in a case.
• Feedback on the verbal description of the imaging finding, not just on the final diagnosis.
• Pre-class material with name pre-training when the topic is dense.

What to avoid in a study group

A group that only projects slides and discusses the final answer trains dual encoding very little. Alternatives:

• One member draws on the board without speaking; the others narrate.
• One narrates without drawing; the others sketch.
• Swap roles each case.
• At the end, each person writes the mechanism sentence in silence and compares.

This low-cost format brings the group closer to the multimedia and retrieval evidence, without needing new software.

When the content is almost entirely verbal

Not everything calls for a figure. Ethical criteria, breaking bad news scripts and some laboratory algorithms live in the verbal channel. Even so, a minimal scheme (a yes/no decision box, a timeline of the conversation) can anchor memory without pretending the topic is anatomical. The coherence rule still applies: if the figure does not teach, do not include it.

For communication, the blog already covers empathy and reasoning and diagnostic uncertainty: there the "drawing" can be the map of the conversation, not an organ.

Closing the method

Dual encoding does not compete with TBL, with cases or with games. It cleans up the micro-design inside any format: what appears together, what is said, what is tested. If your material is already active but you still forget, look first at contiguity, redundancy and retrieval. Only then switch platforms or decks.

How NEXO helps you train

NEXO presents fictional cases with a real clinical presentation and up to six clues. Each clue is verbal; the task of building the mental image of the condition and deciding the next step forces the channels to join. It does not replace atlases or anatomy classes. It complements them by training decisions under partial information, in the same spirit as case-based learning.

See the Economy. Those who coordinate a group can learn about NEXO Closer. Researchers and creators will find NEXO Researcher, NEXO Curator and NEXO Creators.

Frequently asked questions

What is the dual encoding theory?

It is the idea that verbal and nonverbal representations reinforce each other when the material connects them without overloading working memory (Mayer, 2010).

Does this prove that each person has a visual or auditory style?

No. The focus is the design of the task for most learners, not a fixed style label.

Do multimedia principles work in a medical class?

Studies by Issa and colleagues indicate learning gains when classes respect these principles (Issa et al., 2011; Issa et al., 2013).

Is just looking at figures enough?

No. Without a verbal task (naming, explaining, predicting) the gain is limited. Without retrieval testing, retention drops (Larsen et al., 2013; Dunlosky et al., 2013).

How can I apply this in today's study session?

Choose a spatial topic, cover the captions, explain it out loud, redraw it from memory and only then check. Finish with five questions without looking at the material.

What is the difference from active learning?

Active learning is the genre (doing something with the content). Dual encoding is a principle of how to combine channels within that activity (active learning).

Can I use complex animations?

Only if the animation teaches the mechanism step by step. Many medical animations violate basic principles (Yue et al., 2013).

Does NEXO use dual encoding?

NEXO works with textual clues from cases and with the mental construction of the picture; it does not replace illustrated anatomy material.

Play the case of the day on NEXO and, as each clue arrives, sketch mentally what changed in the picture of the condition.

References

1. Mayer RE. Applying the science of learning to medical education. Med Educ. 2010. PMID: [20604850](https://pubmed.ncbi.nlm.nih.gov/20604850/)
2. Issa N, Schuller M, Santacaterina S, et al. Applying multimedia design principles enhances learning in medical education. Med Educ. 2011. PMID: [21752078](https://pubmed.ncbi.nlm.nih.gov/21752078/)
3. Issa N, Mayer RE, Schuller M, Wang E, Shapiro M, DaRosa DA. Teaching for understanding in medical classrooms using multimedia design principles. Med Educ. 2013. PMID: [23488758](https://pubmed.ncbi.nlm.nih.gov/23488758/)
4. Yue C, Kim J, Ogawa R, Stark E, Kim S. Applying the cognitive theory of multimedia learning: an analysis of medical animations. Med Educ. 2013. PMID: [23488757](https://pubmed.ncbi.nlm.nih.gov/23488757/)
5. van Merriënboer JJG, Sweller J. Cognitive load theory in health professional education: design principles and strategies. Med Educ. 2010. PMID: [20078759](https://pubmed.ncbi.nlm.nih.gov/20078759/)
6. Larsen DP, Butler AC, Roediger HL 3rd. Comparative effects of test-enhanced learning and self-explanation on long-term retention. Med Educ. 2013. PMID: [23746156](https://pubmed.ncbi.nlm.nih.gov/23746156/)
7. Raupach T, Andresen JC, Meyer K, et al. Test-enhanced learning of clinical reasoning: a crossover randomised trial. Med Educ. 2016. PMID: [27295475](https://pubmed.ncbi.nlm.nih.gov/27295475/)
8. Dunlosky J, Rawson KA, Marsh EJ, Nathan MJ, Willingham DT. Improving Students' Learning With Effective Learning Techniques: Promising Directions From Cognitive and Educational Psychology. Psychol Sci Public Interest. 2013. PMID: [26173288](https://pubmed.ncbi.nlm.nih.gov/26173288/)
9. Cook DA, Garside S, Levinson AJ, Dupras DM, Montori VM. Instructional design variations in internet-based learning for health professions education: a systematic review and meta-analysis. Acad Med. 2010. PMID: [20520049](https://pubmed.ncbi.nlm.nih.gov/20520049/)
10. Ruiz JG, Mintzer MJ, Leipzig RM. The impact of E-learning in medical education. Acad Med. 2006. PMID: [16501260](https://pubmed.ncbi.nlm.nih.gov/16501260/)
11. Mayer RE, Moreno R. Nine ways to reduce cognitive load in multimedia learning. Educational Psychologist. 2003. (complements Mayer 2010; load principles)
12. Clark JM, Paivio A. Dual-code theory and education. Educational Psychology Review. 1991. (classic synthesis of the Paivio line; complements the PubMed articles above)