Fig 1.
Extracellular multi-unit recording in the anterior optic tract (AOT) of the honeybee brain.
A) Scheme of the bee brain showing the location and the main input and output tracts involving the anterior optic tubercle (AOTU). The AOTU is divided into a major unit and a small lateral ensemble of units. It receives input from the medulla (ME) and lobula (LO) via the prominent AOT (in magenta), which was the target of our neural recordings. Output (in green) to the contralateral AOTU and the lateral complex (LX) is provided by the ventral intertubercle tract (vITT) and the tubercle accessory-lobe tract (TALT), respectively. B) Example of micro-ruby fluorescent staining revealing the place where the electrodes were inserted (arrow). Electrodes were placed in the anterior protocerebral surface at the level of the AOT, thus lateral to the AOTU input region. LA: Lamina; AL: Antennal lobe; CA: Calix of the mushroom body; CB: Central body; scale bars = 50 μm. C-D) Elements of the stimulation and recording set-ups are presented in two different perspectives. C) Two electrodes connected to an amplifier were inserted into the brain of an immobilized bee. The arm pivoting around the bees head was equipped with 3 LEDs (wavelengths: 360 nm, 440 nm, 520 nm) behind a diffuser. D) Two stepper motors controlled the movements of the arms within the semispherical stimulation panorama. An Arduino microcontroller controlled the motors and lightening of the LEDs. One motor controls the movement of the LEDs horizontally (left to right aligns with ventral to dorsal) over a section of 190°. The other motor moves the arm vertically (up to down aligns with anterior to posterior) over a section of 140°. The voltage of the LEDs and the PWM (pulse width modulation) signals were picked up by the same analog to digital converter that recorded the electrophysiological data. The coordinates of the arms (in 266 positions) and the LEDs´ activation were randomized. E) Lateral view of the bee head and right eye during recording. As described by Seidl & Kaiser [50], the middle axis separating the bee eye in its anterior and posterior halves has an inclination angle of 15° in relation to the horizontal plane in which the bee head lays in the holder. F) Representative example of the receptive field of a recorded neuron, calculated considering the bee eye inclination and curvature [50] and the Cassini projection of the 266 positions in which visual stimuli were presented (see section ‘receptive-field projection’). The grayscale represents the recorded spike rate in each coordinate from minimum to maximum values. G) Temporal basis of computing spike rate changes for receptive field analyses. To compute the spike rate changes of the ON-phasic, ON-tonic and OFF response the average spike rate of the before time window (1500 ms before stimulus onset until 10 ms before stimulus onset) was subtracted from the average of the respective time window (ON-phasic: 10 ms—200 ms; ON-tonic: 200 ms—450 ms; OFF: 510 ms—800 ms). Anatomical axes are presented in relation to the bee right eye: do = dorsal, ve = ventral, an = anterior, po = posterior, med = medial, lat = lateral.
Table 1.
Distinct chromatic categories of neurons recorded in the honeybee AOT.
Fig 2.
Broad-band and narrow-band units recorded in the AOT.
Average peristimulus time histograms (PSTH) of 266 presentations are shown above spike raster plots across four presentations of each spectral light (UV, blue and green). A: Broad-band ON-tonic unit; B: Broad-band ON-phasic unit; C: Broad-band ON-OFF unit; D: Narrow-band unit. Each example of single unit shown in A-D come from a distinct bee. Timing of light presentation (= 500 ms) is indicated by the black bar. First spike latency (FSL; mean ± SD) in response to each stimulus is shown at the right of each respective raster plots.
Fig 3.
Color-opponent units recorded in the AOT.
Average peristimulus time histograms (PSTH) of 266 presentations are shown above spike raster plots across four presentations of each spectral light (UV, blue and green). A: UV+ / B- / G- unit; B: UV(s)/B(f)/G(f) unit; C: UV(f)/B(s)/G(s) unit; D: UV(on) / B(on-off) / G(on-off) unit. Each example of single unit shown in A-D come from a distinct bee. Timing of light presentation (= 500 ms) is indicated by the black bar. First spike latency (FSL; mean ± SD) in response to each stimulus is shown at the right of each respective raster plots.
Fig 4.
Receptive field (RF) structure of broad-band (BB) ON-phasic (A), broad-band ON-OFF (B) and narrow-band (NB) AOT neurons (C). A and C: RF activation during ON-response phase and average PSTH are shown for each spectral stimulus (UV, blue and green). Black bars indicate the stimulus presentation of 500ms. B: Equivalent to A and C, but RF activation is also shown during the OFF-response phase for a BB ON-OFF neuron. Each example shown in A-C is from a different bee. Light gray indicates regions of the RF where light stimuli failed or were unreliable. D: Average RF sizes recorded for UV, blue and green in BB (left; N = 26 units recorded in 22 bees) and NB (right; N = 12 units recorded in 12 bees) neurons. Short and long axes of the RF are indicated by horizontal and vertical stripes within boxplots, respectively. The white square inside the box shows the mean. The boundaries of the box indicate the 25th and 75th percentile. Whiskers indicate the minimum and maximum values. E: Overlap between RFs for each pair of spectral stimuli (UV vs blue; UV vs green, blue vs green) in BB ON-phasic (N = 11 units; 10 bees), BB ON-OFF (N = 12 units; 11 bees) and NB neurons (N = 12 units; 12 bees). The dots at the right of bars show individual plots of each independent unit. Different letters indicate significant differences (p < 0.05) in Wilcoxon matched-pairs tests.
Fig 5.
Receptive field (RF) structure of a UV+/B-/G- (A) and a UV(s)/B(f)/G(f) color-opponent AOT neuron (B). RFs expressed for excitation and/or inhibition are shown during the ON-phasic (left), the ON-tonic (middle) and the OFF (right) responses. Average PSTHs of 266 presentations are presented below the RFs for each spectral stimulus (UV, blue and green). Black bars indicate the stimulus presentation of 500ms. Examples shown in A and B were recorded in distinct bees. Light gray indicates regions of the RF where light stimulation failed or appeared to be unreliable. C: Average RF sizes recorded for UV, blue and green in color opponent neurons (N = 34 neurons recorded in 29 bees). Short and long axes of the RF are indicated by horizontal and vertical stripes within boxplots, respectively. The white square inside the box shows the mean. The boundaries of the box indicate the 25th and 75th percentile. Whiskers indicate the minimum and maximum values. D: Overlap between RFs for each pair of spectral stimuli (UV vs blue; UV vs green, blue vs green) in UV+/B-/G- color-opponent neurons (N = 13 neurons; 13 bees). The dots at the right of bars show individual plots of each neuron recorded. Excitatory RF overlap is represented by gray and inhibitory RF overlap by black bars and dots. Different letters indicate significant differences (p < 0.05) in Wilcoxon matched-pairs tests.
Fig 6.
Receptive fields of some color-opponent AOT units with both double spectral and spatial opponency.
Regions of the RF presenting spatial opponency during their on-tonic response phase are shown for a UV+/B-/G- unit (A) and a UV(s)/B(f)/G(f) unit (B), as indicated by dark red or dark blue on the overlapped RFs for each color pair: UV vs green, UV vs blue, blue vs green. Dark red indicates regions excited by the first and inhibited by the second color of each pair, whereas dark blue indicates regions inhibited by the first and excited by the second color of each pair. The units shown in (A) and (B) correspond to the ones presented in Fig 5A and 5B, respectively. They both presented large regions of their on-tonic RFs that were excited by UV and inhibited by blue or green. Some smaller RF regions were excited by blue and inhibited by green or vice-versa, during the on-tonic period (A-B). C: Another UV+/B-/G- unit presented RF regions with spatial opponency during all three response periods: On-phasic, on-tonic and off-phasic. As observed for the other color-opponent units (A-B), spatial opponency was broader during the on-tonic phase for UV vs blue or green (dark red). Light gray indicates regions of the RF where light stimulation failed or appeared to be unreliable. Dark gray regions of the RF did not present spatial opponency properties. D: Proportion of the RF presenting spectral and spatial opponency during the on-tonic response in UNIT A (A) and UNIT B (B) for each color pair. E: Proportion of the RF presenting spectral and spatial opponency during all response phases in UNIT C (C) for each color pair. Double opponency, as revealed by the simultaneous presence of red and blue bars in a same response phase, was detected during the on-tonic responses of UNITs B and C, as well as the off response phase of UNIT C. Each of the three example units (A, B and C) were recorded in a different bee.
Fig 7.
Responses to a mixture of UV, blue, green light (bee´s white) in comparison to the single spectral components reveal hipoadditive (A), suppressive (B) or synergic (C) combinatory effects. Box plots represent the spike rates of each example neuron (A, B and C) from values recorded in three presentations of each stimulus over six distinct RF coordinates. Each of the three example units (A, B and C) were recorded in a different bee. The horizontal line within the box shows the median. The boundaries of the box indicate the 25th and 75th percentile. Whiskers indicate the minimum and maximum values. The dots at the right of each box show data dispersion (N = 18 presentations per stimulus). Different letters indicate significant differences (p < 0.05) in Wilcoxon matched-pairs tests. D: Types of neurons displaying each combinatory effect.
Fig 8.
Broad-band (BB) neurons in response to chromatic combinations.
A-C: Responses of a BB ON-phasic neuron to UV (A), blue (B) and green (C) stimulus over distinct backgrounds. Arrows indicate OFF-responses that emerged after chromatic combination. D-F: Responses of a BB ON-OFF neuron to UV (A), blue (B) and green (C) stimulus over distinct backgrounds. Bins on the abscissa indicate 100ms time intervals. Horizontal bars indicate background (700ms) and stimulus (500ms) presentation. Different letters indicate significant difference between spike rate during the onset of each spectral light presented alone or in the presence of two distinct spectral backgrounds (Friedman ANOVA; A-B and D-F: Chi2 > 31.71, p < 0.01; Wilcoxon test, A-F: p < 0.02 for all significant comparisons).
Fig 9.
Color-opponent neurons in response to chromatic combinations.
A-C: Responses of a UV+/B-/G- neuron to UV (A), blue (B) and green (C) stimulus over distinct backgrounds. D-F: Responses of a UV(on)/B(on-off)/G(on-off) neuron to UV (A), blue (B) and green (C) stimulus over distinct backgrounds. Bins on the abscissa indicate 100ms time intervals. Horizontal bars indicate background (700ms) and stimulus (500ms) presentation. Different letters indicate significant differences between spike rates during the onset of each spectral light presented alone or in the presence of two distinct spectral backgrounds (Friedman ANOVA; A-F: Chi2 > 25.56, p < 0.01; Wilcoxon test, A-F: p < 0.03 for all significant comparisons).
Fig 10.
Three types of spectral dominance recorded in AOT neurons of the honeybee, and their associated inhibitory pathways, as revealed by responses to chromatic combinations.
A) Broad-band (BB) ON-phasic neurons present similar excitatory responses to each spectral stimulus (upper curves). However, responses to chromatic mixtures reveal complete domination of longer wavelengths over shorter wavelengths (lower curves), suggesting inhibitory pathways (black arrows) that produce this suppressive phenomena. These neurons usually present smaller receptive fields (RFs) than BB ON-OFF or any color-opponent (CO) neuron type. B) CO UV+/B-/G- neurons respond with tonic excitation to UV and tonic inhibition to blue or green (upper curves). While a background of longer wavelength (e.g. green) inhibits (solid arrows) responses to any shorter wavelength (e.g. UV) added to the mixture (lower curves), a blue background also partially inhibits responses to green (dashed arrow). C) BB ON-OFF and CO UV(on)/B(on-off)/G(on-off) neurons while different in their responses to individual spectral stimuli (upper curves) presented very similar responses to chromatic mixtures (lower curves), with total inhibition of shorter wavelengths by longer wavelengths (solid arrows) and partial inhibition of longer wavelengths by a background of shorter wavelength. Therefore, ‘BB’ neurons as the ones illustrated in (A) and (C) are also connected to color-opponent circuits. Neuronal categories presenting OFF responses to single spectral stimuli and partial inhibition (dashed arrows) of shorter over longer wavelengths (B-C) have larger RFs than ON-phasic neurons (A).