After the Photograph Disappears: Spatial Traces of Looking While Drawing

When someone asks you to sketch a photograph from memory, your hand begins working on a surface that does not contain the original scene. Your eyes cannot return to the photograph to check a missing detail. Yet the way you explored that photograph may still have a relationship with where you look while drawing. The question is not simply whether the finished sketch resembles the scene. It is whether something specific about your earlier visual exploration remains detectable after the scene has disappeared.

Stephanie Huette, Christopher T. Kello, Theo Rhodes, and Michael J. Spivey investigated this problem in a study published in 2013. They recorded eye and pen movements during a scene-viewing and drawing task. Their experiment did not compare drawing with writing as competing methods for improving memory. It examined coordination between looking and subsequently drawing. Two possible kinds of correspondence proved important: visiting locations in the same temporal sequence and concentrating activity in similar broad spatial regions. Those are related ideas, but they make different predictions. [1]

Three records from one task

Sixteen university students participated. Four were excluded because the eye tracker could not be calibrated accurately enough, and one because the drawing task was not performed properly, leaving eleven participants in the analyses. None reported being an expert artist. Each participant saw six complex natural-scene photographs in random order. A photograph remained available for thirty seconds. After a four-second preparation interval, participants had ninety seconds to draw it without the original image being visible. [1]

The apparatus matters for interpreting the behavior. Participants drew with a pen on a digitizing tablet that was separate from the computer monitor. Their black-and-white lines appeared on the monitor, where their eyes could follow them. This was not the same arrangement as looking directly at a pen moving over paper. The hand and arm did not cover the displayed drawing. Any application of the findings to ordinary sketching should keep that difference in view. [1]

The researchers obtained three position records: the eyes while studying a photograph, the eyes while drawing, and the pen while drawing. They removed unusable eye samples, including blinks and off-screen positions. The pen record included only periods when the pen contacted the tablet and produced lines. Expressing these records in corresponding coordinates made comparisons possible. A position record still does not reveal everything a participant remembered or intended, however. It measures where activity occurred, rather than providing a direct picture of the entire mental representation. [1]

Coordination during drawing was not the same as replay

The first analysis concerned eye and pen activity within the drawing period. Their locations tended to be closer together at approximately the same time than at widely separated times. This supported coordination during the ongoing task. On average, the temporal pattern did not imply that the eyes always commanded the hand from a fixed position ahead of it. Eyes could lead or follow the pen. Coordinated behavior need not have one component acting as the permanent leader. [1]

The memory question required a different comparison: could the thirty-second study trajectory be matched to activity during the ninety-second drawing period? The researchers stretched the study time axis by a factor of three to make the two periods comparable in length. If drawing involved a straightforward slower replay of the earlier visual exploration, matching locations might become evident at particular time relationships. That kind of temporal correspondence was not supported. Coordination within drawing and a temporally preserved repetition of earlier looking were not interchangeable findings. [1]

This negative result did not establish that the study experience had left no trace. The test addressed a particular form of alignment. A more complex transformation of timing, or a correspondence preserved only in spatial organization, might not emerge from it. The result therefore counted against a simple replay account, rather than demonstrating the complete absence of temporal information in memory. The next analysis changed the question from when locations were visited to where activity was distributed. [1]

Changing the scale of comparison

The researchers adapted an approach based on the Allan factor to compare spatial clustering at different scales. In broad terms, they divided the image area into square tiles of different sizes and examined the distribution of recorded positions. Small tiles emphasize fine distinctions; larger tiles emphasize broader regions. This was not an aesthetic rating of how convincing a sketch looked. It was a way to examine the spatial organization of eyes and pen without requiring every visit to occur in the same order or last the same amount of time. [1]

Consider an invented illustration: you might spend much of your study time exploring a bridge on the right side of a landscape. While sketching later, your eyes might repeatedly return to that general region without retracing every earlier fixation in sequence. The example explains the distinction; it is not an additional observation from the experiment. Spatial correspondence can retain some structure while allowing the order of actions to change. In the study, correspondences across study and drawing were more apparent at larger spatial scales than at fine ones. [1]

Even that similarity needs a suitable baseline. Many people might tend to look toward the center of a screen regardless of the photograph. Alternatively, a prominent feature in a particular image might draw everyone's attention to a similar region. Neither possibility requires a highly specific trace of one person's earlier exploration. The researchers therefore compared the original pairings with two kinds of surrogate pairings. One paired records from the same participant but different images. The other paired records for the same image but different participants. [1]

The first comparison addressed general patterns characteristic of a person. The second addressed patterns shared because people were viewing the same scene. Correspondence exceeding those baselines suggested that some organization was specific to both the person and the image. That is a more demanding test than noticing that two trajectories occupy the same obvious part of a photograph. It asks whether the original combination carries information beyond those general sources of similarity. [1]

An important exception limits the conclusion. The paper's results section explicitly reports that the comparison between study-period eyes and drawing-period pen did not show a significant difference from baseline. The evidence must not be compressed into a claim that every earlier gaze path was directly reproduced by the hand. The relationship between study gaze and drawing gaze supplies a firmer basis for the narrower argument. Different movement pairings did not all yield equally strong support. [1]

A spatial trace is not a complete theory of memory

The findings allow a more selective picture of remembering than either perfect internal photography or a complete disappearance of earlier action. Broad spatial organization can remain detectable even when a simple temporal replay is not. Weak correspondence at fine scales does not, by itself, prove that every fine detail was forgotten. Pen movement characteristics, recording resolution, and measurement error can also affect which correspondences become measurable. Likewise, the authors' interpretation does not exclude visual and symbolic forms of representation operating alongside perceptual-motor information. [1]

The limits of the experiment are substantial and useful. Eleven students drawing six photographs do not establish how every person remembers every scene. This design did not manipulate gaze to show that making prescribed eye movements improves memory. It did not test the benefits of art training, long-term retention, or a clinical intervention. Its results concern correspondence during a particular task, not a general technique for strengthening recall.

Within that scope, the evidence supports an interesting conclusion. Drawing from memory can connect the spatial organization of earlier visual exploration with later behavior, rather than relying only on coordination occurring at the instant the pen moves. The photograph is gone, but some aspects of how it was explored can remain apparent. What returns is not a complete, unaltered playback of looking. It is a more limited spatial relationship, identified by asking separately about timing, scale, and the similarities that would have occurred anyway.

Sources

[1] Huette, S., Kello, C. T., Rhodes, T., and Spivey, M. J. (2013). Drawing from Memory: Hand-Eye Coordination at Multiple Scales. PLOS ONE, 8(3), e58464. https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0058464 . DOI: 10.1371/journal.pone.0058464. Original open-access research under a Creative Commons Attribution license. This article is an independently written explanation and assessment of that study.