Setups for studying individuality walking flies.

a Overview of a classical (left, fluorescent tubes) and LED (middle) Buridan assay (behavioral platform diameter = 120mm). A MATLAB script provides single fly tracking and analysis of behavioral key components (right). b-d persistence of individual walking patterns across time (b), different stripe numbers (c), and changes in background illumination (d).

Individual traits of walking flies persist over time but depend on the visual scenery.

a-c Correlation of key behavioral parameters over time (a), different stripe numbers (b), and changes in background illumination (c). Correlation coefficients are displayed in the upper left corner of each plot. *=p<0.05, **=p<0.01, ***=p<0.001, for detailed fly numbers per experimental condition, see supplemental Fig. S2.

A high-throughput assay for studying the influence of temperature, illumination, and arena shape on individuality.

a Setup overview. Flies are filmed from below against a near-infrared backlight and are walking within a 5×5 array of small Buridan chambers (b) or one of up to 144 Y-maze arenas (c). The setup is housed within a temperature- and humidity-controlled enclosure (600 x 450 x 820 mm). The provided MATLAB code corrects lens distortions, automatically detects the arenas, and tracks each fly’s position over time, allowing for generating high-throughput single-fly behavioral data (d).

The influence of temperature, illumination, and arena shape on individuality in walking flies.

A Individuality in walking patterns across different temperatures, illumination, and arena type. Diameter circular arena = 49mm, arm length from center Y-maze = 13.5mm. b changes in the % of time walked for all tested flies depending on environmental conditions. c Changes in the % of time walked grouped by temperature, illumination, and arena. d Correlation of key behavioral parameters across temperature, illumination, and arena type. For detailed fly numbers per experimental condition, see supplemental Fig. S6.

A virtual flight simulator for studying individuality in flying flies.

a Setup overview (225 x 225 x 600 mm). Flies are glued to a steel pin and placed in a magneto-tether surrounded by an LED matrix for stimulus presentation. b Close-up photo of a tethered flying fly. c Schematic setup overview. Flies are filmed from below through an IR-pass filter against near-infrared back illumination. A connected pump provides air puffs to reinitiate flight if necessary. We provide a mounting port for an optional IR-laser for heat punishment via a dichroic mirror. d A MATLAB script controls stimulus presentation (optomotor or Buridan stimulus) with an angular extent of 360° azimuthally and 100° vertically and provides simultaneous real-time tracking of flight-heading over 360° (e). Before an experiment, each fly’s ability to rotate was tested using a 120°/s CW or CCW stripe rotation (optomotor stimulus), resulting in measured rotational flight movements close to the rotational velocity of the stripe pattern (e, upper left) and uniform distribution of angles (e, lower left). This was followed by each fly flying under the Buridan stimulus for 10min (d, right) and an analysis of heading choices (e, upper right) and their angular distribution (e, lower right).

Individual traits of flying flies persist over time but depend on the visual scenery.

a Angular histograms show the persistence of individual heading choices over consecutive days and across different contrast ratios. b,c Correlation of key behavioral parameters over time (b) and depending on changes in visual contrast (c). For detailed fly numbers per experimental condition, see supplemental Fig. S10.

Visual attention persists across different behavioral states.

a Experimental overview. Flies were untethered after flying within the flight simulator for 10 min (left) and walking for 15 min (right) under the same Buridan stimulus (stripe size=11°). Single fly data (middle) and population data (bottom) reveal stripe fixation behavior when walking and during flight. b Correlation of key behavioral parameters depending on the behavioral state (flight=x-axis, walking=y-axis). For detailed fly numbers per experimental condition, see supplemental Fig. S12.

A hierarchy of environmental and behavioral contexts based on their influence on individuality.

a Overview summarizing correlation coefficients (p<0.05) for key behavioral parameters depending on behavioral context, time, temperature, and visual input. b Quantification of the influence of behavioral and environmental contexts on the persistence of individuality.

Construction of Buridan assays. a Two different types of Buridan assays (fluorescent tubes left, flexible LED matrix right) and the parts necessary for construction. Part numbers are referenced in supplemental table T2 and for 3D printed parts under github.com/LinneweberLab/individuality-assays for replication. Black numbers indicate off-the-shelve parts; blue numbers indicate 3D-printed parts. b construction of Buridan bottom plates. An 850nm LED strip is wrapped around a translucent acrylic disk (11), providing homogeneous near-infrared background illumination from below. c schematic illustration of the tracking procedure. The provided MATLAB script automatically reads video frames, tracks the X- and Y-coordinates of individual flies over time, and saves data for further analysis.

Population responses over time, varying stripe numbers and background illumination. a-c Boxplots illustrating key behavioral parameters for the whole population (blue) and separately for females/males (green/magenta) over time (a), varying stripe numbers (b) and background illumination (c). Animal numbers are indicated next to each plot.

Sex-specific differences in individual traits. a-d Correlation of key behavioral parameters for females and males over time (a), different stripe numbers (b), changes in background illumination (c), and depending on the usage of a clear dome around the arena (d). Magenta=females, green=males. Correlation coefficients are displayed in the upper left corner of each plot. *=p<0.05, **=p<0.01, ***=p<0.001, for detailed fly numbers per experimental condition, see supplemental Fig. S2.

Construction of a high-throughput assay for studying the influence of temperature, illumination, and arena shape on individuality. a Overview of parts necessary for constructing the assay. Part numbers are referenced in supplemental table T2 and for 3D printed parts under github.com/LinneweberLab/individuality-assays for replication. Black numbers indicate off-the-shelve parts; blue numbers indicate 3D-printed parts. b Schematic illustration of the tracking procedure. The provided MATLAB script automatically reads video frames, detects single arenas, tracks the X- and Y-coordinates of individual flies over time, and saves data for further analysis. c Exemplary temperature- and humidity curves were measured within the center arena during heat-up to 25 and 30°C, respectively. Stable plateaus are reached after ∼20min. d The camera was calibrated using the MATLAB calibration toolbox. The provided MATLAB script uses these camera parameters to correct lens distortion artifacts during tracking. e Screenshots of the tracking procedure using the multi-Buridan assay. Arenas are detected automatically, and individual flies get thresholded after background subtraction, resulting in reliable trajectories of each individual fly. f Same as e, but for the Y-maze assay. The box on the right shows an enlarged depiction of a single arena. g Same as e, but for linear assays, demonstrating the usability of the setup with other arena shapes. The box on the right shows an enlarged depiction of a single arena.

Preliminary data of a fly population (TopBanana) illustrating orientational responses. Flies walk between the 2 stripes and show randomly distributed headings in darkness, independent of temperature. a Velocity- (left, transparency added to tracks), transition- (middle), and angular plots (right, angles from moving periods only, non-axial) for vertical stripes positioned at 0° (top), 90° (middle) or in complete darkness (bottom), respectively for all tested flies at 25°C. Scale bar = 10mm. b) Same as a) but at 30°C. Fly numbers are indicated in the velocity plots.

Sex- and genotype-specific population responses over varying temperature, illumination and arena shape in walking flies. a Boxplots illustrating key behavioral parameters for the whole tested CantonS population (left, blue=23°C, red=32°C) and separately for females/males (green/magenta) over varying temperature, illumination and arena shape. Animal numbers are indicated next to each plot. b Same as a, but for all tested TopBanana flies.

Sex- and genotype-specific differences based on the influence of temperature, illumination, and arena shape on the persistence of individual traits in walking flies. a Correlation of key behavioral parameters for females and males depending on temperature, illumination, and arena. Magenta=females, green=males. Correlation coefficients are displayed in the upper left corner of each plot. *=p<0.05, **=p<0.01, ***=p<0.001, for detailed fly numbers per experimental condition, see supplemental Fig. S6. b same as a but sorted by genotype (red=CantonS, blue=TopBanana).

Complete correlation matrix across all behavioral parameters and contexts. Correlation coefficients with p<0.05 are color coded.

Construction of a virtual flight simulator for studying the persistence of individual traits in flying flies. a Overview of parts necessary for constructing the assay. Part numbers are referenced in supplemental table T2 and for 3D printed parts under github.com/LinneweberLab/individuality-assays for replication. Black numbers indicate off-the-shelve parts; blue numbers indicate 3D-printed parts. b schematic illustration of the tracking procedure. The provided MATLAB script works in real-time at ∼90Hz on a 4GHz processor. The code automatically reads camera frames, computes the heading angle over time, and saves data for further analysis. c Illustration of the tracking procedure. The camera frame is thresholded using 2 different Gaussian filter settings to separately threshold the body (yellow) and legs/wings (dark blue). An ellipse is fitted around the thresholded body pixels, and its main axis is used to measure the distribution of body pixels within 4 virtual quadrants whose origin is the centroid of the fly. The head region can be determined because it appears smaller in the camera images than the abdominal region of the fly.

Population responses over time and different visual contrasts in flying flies. Boxplots illustrating key behavioral parameters for the whole population (blue) and separately for females/males (green/magenta) over time and varying stripe contrasts (b). Animal numbers are indicated next to each plot.

Sex-specific differences of individual traits in flying flies persist over time but depend on the visual scenery. a Correlation of key behavioral parameters for females and males over time and dependent on stimulus contrast. Magenta=females, green=males. Correlation coefficients are displayed in the upper left corner of each plot. *=p<0.05, **=p<0.001, ***=p<0.0001, for detailed fly numbers per experimental condition, see supplemental Fig. S10.

Population responses of orientational parameters across different behavioral states. Boxplots illustrating orientational parameters under different behavioral states for the whole test population (blue) and separately for females/males (green/magenta). Animal numbers are indicated next to each plot.

Attention persists across different behavioral states in males but not females. a Correlation of key behavioral parameters for females and males depending on the behavioral state. Magenta=females, green=males. Correlation coefficients are displayed in the upper left corner of each plot. *=p<0.05, **=p<0.001, ***=p<0.0001, for detailed fly numbers per experimental condition, see supplemental Fig. S12.

Description of automatically computed output parameters

Part list for setup replication