Olfactory habituation and heterosensory and homosensory dishabituation.

A) The model of short exposure olfactory habituation (Semelidou, Acevedo et al. 2018), establishing that habituation requires inhibition of the LH-driven innate avoidance response to OCT. Model created with BioRender.com/0qwwz1v. B-G) Mean Performance Indices ± SEM are shown in all panels. The naïve avoidance response (AVD) is shown by the white bar for OCT and white striped bar for yeast odor (YO). The habituated response (HAB) is shown by the grey bars for OCT and the striped, grey bar for YO. The dishabituated response is shown in magenta for the footshock dishabituator, in orange for YO and in cyan/petrol for OCT. Stars indicate significant differences (p<0.005) from the habituated response and daggers significant difference (p<0.005) from naïve avoidance. All statistical details are presented in Supplemental Table 1. B,C) Naïve responses of Canton-S and w1118 flies to OCT (AVD) and habituated (HAB) responses to OCT, dishabituated either by a single 45V electric shock (bolt) or a puff of yeast odor (plume). n=14 for B and n=12 for C.

D) Naïve response of w1118 and Orco2 homozygotes to OCT and YO. Daggers indicate significant differences between the respective naïve responses to OCT and YO of Orco2 homozygotes controls. n=12 E) Footshock habituation cannot be dishabituated by YO presentation in Orco2 homozygotes, unlike in control animals. n=16. F) Time-dependent changes in the response of w1118 flies to YO. Daggers indicate significant differences (p<0.005) between the responses after 10 and 90 sec compared to the shortest exposure (5 sec) to YO. n=13 G) Time-dependent changes in the response of w1118 flies to OCT. Daggers indicate significant differences (p<0.005) between the response after 90 sec compared to the shorter exposures (5 and 10 sec) to OCT. n=14 H) Habituation of control flies to YO and dishabituation by OCT presentation. n=16.

The mushroom bodies are essential for homosensory and heterosensory dishabituation.

A) Schematic of the MB illustrating its tripartite structure. Created with BioRender.com/sc2j0u0. B-F) Mean Performance Indices ± SEM are shown in all figures. The naïve response is shown in white bars and is always significantly different from the habituated response (grey bars). Heterosensory (footshock) dishabituation is indicated by the bolt and homosensory (YO), by the plume. Exposure to blue light is indicated by the LED bulbs and increasing intensity (as indicated by the Volts) by deepening shades of blue. Stars in all Figures indicate significant differences (p<0.005), from the habituated response and daggers significant difference (p<0.005) from naïve avoidance. All statistical details are presented in Supplemental Table 1. B) Functional silencing of the MB neurons (IN) with UAS-shibirets under dncGal4 resulted in the expected habituation after 4 min of OCT exposure, but no dishabituation after stimulation with a 45V electric shock (magenta bar) or YO (orange bar). At the permissive conditions (UN), habituation and both types of dishabituation were not affected. n=14 C) Flies expressing ChR2XXL in their MBs avoid, habituate and dishabituate with footshock or YO normally. Photoactivation of MB neurons in OCT-habituated flies with three different light intensities of 4.5V, 5V and 6V, restored OCT avoidance at all intensities. n=16 D) Photoactivation of MB neurons in naïve flies did not affect OCT avoidance. n=16. E) Control dncGal4>+ and UAS-shibirets>+ flies were subjected to the same induction conditions as the experimental group and did not present habituation or dishabituation impairments. n=14. F) Control dncGal4>+, not expressing ChR2XXL did not dishabituate upon photoactivation eliminating the possibility of dishabituation via visual stimulation. n=16. G) Blue light exposure at all intensities used did not affect OCT avoidance in naïve flies. n=16.

Distinct MBns regulate olfactory habituation and dishabituation.

Except for the schematics, mean Performance Indices ± SEM are shown in all figures. The naïve response is shown in white bars and is always significantly different from the habituated response (grey bars). Heterosensory (footshock) dishabituation is indicated by the bolt and homosenssory (YO), by the plume. Exposure to blue light is indicated by the LED bulbs and increasing intensity (as indicated by the Volts) by deepening shades of blue. Stars in all Figures indicate significant differences (p<0.005), from the habituated response and daggers significant difference (p<0.005) from naïve avoidance. All statistical details are presented in Supplemental Table 1. A) Schematic of the axonal projections (lobes) of αβ MBns highlighted in light red. B) Functional silencing (IN) of the αβ neurons with UAS-shibirets under 17dGal4 compared to control animals carrying, but not expressing the transgene (UN). Dishabituation with a 45V electric shock (magenta) and YO (orange), remained unaffected in both conditions. n=12. Additional controls run in parallel are presented in Supplemental Table 1. C) Photoactivation of αβ neurons with 3 light intensities in OCT-habituated flies restored avoidance of the odorant. n=14. D) Photoactivation of αβ neurons in naïve flies with 3 light intensities did not affect OCT avoidance. n=14. E) Schematic of the axonal projections (lobes) of α΄β΄ MBns highlighted in purple. F) Functional silencing (IN) of α΄β΄ neurons with UAS-shibirets under VT030604Gal4 compared to control animals carrying, but not expressing the transgene (UN). Avoidance, habituation and dishabituation with a 45V electric shock (magenta) or YO (orange) were not significantly different upon silencing α΄β΄ MBns and were also unaffected in the UN condition. n=14. Additional controls run in parallel are presented in Supplemental Table 1. G) Photoactivation of α΄β΄ neurons under VT030604Gal4 in OCT-habituated flies with 3 light intensities did not restore avoidance, while a single footshock and YO exposure did. n=12. H) Photoactivation of α΄β΄ MBns in naïve flies resulted in light intensity dependent decrement in OCT avoidance. n=12. I) Schematic of the axonal projections (lobes) of γ MBns highlighted in blue. J) Functional silencing (IN) of γ neurons with UAS-shibirets under VT44966Gal4 compared to control animals carrying, but not expressing the transgene (UN). Dishabituation with a 45V electric shock (magenta) and YO (orange), remained unaffected in both conditions. n=12. Additional controls run in parallel are presented in Supplemental Table 1. K) Photoactivation of γ neurons under VT44966Gal4 in OCT-habituated flies with 3 light intensities reinstated avoidance in light intensity dependent manner. n=14. L) Photoactivation of γ MBns in naïve flies did not affect OCT avoidance. n=14. Models created with BioRender.com/d6ux825.

Dopaminergic Neurons drive Heterosensory and Homosensory dishabituation in Drosophila.

Except for the schematics, mean Performance Indices ± SEM are shown in all figures. The naïve response is shown in white bars and is always significantly different from the habituated response (grey bars). Heterosensory (footshock) dishabituation is indicated by the bolt and homosenssory (YO), by the plume. Exposure to blue light is indicated by the LED bulbs. Stars in all Figures indicate significant differences (p<0.005), from the habituated response and daggers significant difference (p<0.005) from naïve avoidance. All statistical details are presented in Supplemental Table 1. A,B,C) Schematics of all Dopaminergic neurons (DANs), the PAM-DAN cluster and the PPL1 DAN cluster, innervating the MB. Created with BioRender.com/pv1ckab. D) Functional silencing of all DANs (IN), with UAS-shibirets under pleGal4 did not affect habituation after 4 min of OCT exposure, but dishabituation after stimulation with a 45V electric shock (magenta bar) and YO (orange bar) were impaired. At the permissive conditions (UN), habituation and both types of dishabituation were not affected. n=16. Additional controls run in parallel are presented in Supplemental Table 1. E) Photoactivation of all DANs in OCT-habituated flies with three different light intensities restored OCT avoidance in light intensity dependent manner. At 4.5 V avoidance was no different from the habituated performance (p=0.3449). At 5V, while significantly different from the habituated performance (p<0.0001) it was also different (p=0.0069) from the naïve avoidance. marginally n=16 F) Photoactivation of all DANs in naïve flies without any exposure to OCT did not affect OCT avoidance. n=14. G) Functional silencing (IN) of PAM-DANs with UAS-shibirets under 0273Gal4 was permissive to habituation after 4 min of OCT exposure and shock dishabituation (magenta bar), but no dishabituation after stimulation with YO (orange bar). At the permissive conditions (UN) YO dishabituation was not affected. n=16. Additional controls run in parallel are presented in Supplemental Table 1. H) Photoactivation of PAM-DANs in OCT-habituated flies restored OCT avoidance only at the higher light intensity. n=10. I) Photoactivation of PAM-DANs in naïve flies without any exposure to OCT did not affect OCT avoidance. n=12. J) Functional silencing (IN) of PPL1-DANs with UAS-shibirets under MB504BGal4 did not affect habituation after 4 min of OCT exposure and dishabituation with YO (orange bar), but dishabituation after stimulation with a 45 V electric shock (magenta bar) was impaired. At the permissive conditions (UN) shock dishabituation was not affected. n=16. Additional controls run in parallel are presented in Supplemental Table 1. K) Photoactivation of PPL1-DANs in OCT-habituated flies restored OCT avoidance only at the lower light intensity. n=16. L) Photoactivation of PPL1-DANs in naïve flies without any exposure to OCT did not affect OCT avoidance. n=16.

The role of PPL1 neuronal subsets in heterosensory dishabituation.

Mean Performance Indices ± SEM are shown in all figures. The naïve response is shown in white bars and is always significantly different from the habituated response (grey bars). Heterosensory (footshock) dishabituation is indicated by the bolt and homosensory (YO), by the plume. Exposure to blue light is indicated by the LED bulbs. Stars in all Figures indicate significant differences (p<0.005), from the habituated response and daggers significant difference (p<0.005) from naïve avoidance. All statistical details are presented in Supplemental Table 1. A,D) Schematics of PPL1-γ2α΄1 and PPL1-γ1pedc neurons. B) Photoactivation of PPL1-γ2α΄1 neurons under MB296BGal4 in OCT-habituated flies restored OCT avoidance at the low, but not the stronger light intensities (p=0.0087 versus the habituated response and p=0.0162 versus naïve avoidance). n=18 C) Photoactivation of PPL1-γ2α΄1 neurons in naïve flies resulted in light intensity dependent decrement in OCT avoidance from naïve avoidance (p=0.0064 at 5V and p<0.0001 at 6V). n=17. E) Photoactivation of PPL1-γ1pedc neurons under MB320CGal4 in OCT-habituated flies restored OCT avoidance at the low and strong light intensities respectively. n=16. F) Photoactivation of PPL1-γ1pedc neurons in naïve flies resulted in light intensity dependent decrement in OCT avoidance with performance at naïve levels only at 4.5V (p=0.6747). n=12. G,J) Schematics of PPL1-α2-α΄2 and PPL1-α3 neurons. H) Photoactivation of PPL1-α2-α΄2 neurons under MB058BGal4 in OCT-habituated flies with 3 light intensities did not restore avoidance, while the response recovered after a single footshock and YO exposure. n=14 for H. I,L) Photoactivation of PPL1-α2-α΄2 and PPL1-α3 neurons in naïve flies with 3 light intensities did not affect OCT avoidance. (n=12 for I and n=14 for L). K) Photoactivation PPL1-α3 neurons under MB630BGal4 in OCT-habituated flies with 3 light intensities did not restore avoidance, while the response recovered after a single footshock and YO exposure. n=16. L) Photoactivation of PPL1-α3 neurons in naïve flies with 3 light intensities did not affect OCT avoidance. n=14. Models created with BioRender.com/u8rdkt1.

APL and other GABAergic neurons regulate homosensory Dishabituation.

Mean Performance Indices ± SEM are shown in all figures. The naïve response is shown in white bars and is always significantly different from the habituated response (grey bars). Heterosensory (footshock) dishabituation is indicated by the bolt and homosensory (YO), by the plume. Exposure to blue light is indicated by the LED bulbs. Stars in all Figures indicate significant differences (p<0.005), from the habituated response and daggers significant difference (p<0.005) from naïve avoidance. All statistical details are presented in Supplemental Table 1. (A) Schematic of APL neurons innervating the MB. Created with BioRender.com/4z4gbrz. (B) Functional silencing of APL neurons with UAS-shibirets under APLGal4 was permissive to habituation after 4 min OCT exposure and shock dishabituation (magenta bar), but dishabituation after YO exposure (orange bar) was impaired. At the permissive conditions (UN) YO dishabituation was not affected. n=12. Additional controls run in parallel are presented in Supplemental Table 1. (C) Photoactivation of APL neurons in OCT-habituated flies restored OCT avoidance only at the higher light intensity. n=14. (D) Photoactivation of APL neurons in naïve flies with 3 light intensities did not affect OCT avoidance. n=14. (E) Abrogation of the GABAergic output of APL neurons with UAS-GAD-RNAi under APLGal4;Gal80ts at 30°C did not affect habituation after 4 min OCT exposure or dishabituation with shock (magenta bar), but dishabituation with YO stimulation was impaired (orange bar). At the permissive conditions (UN) dishabituation with YO was not affected. n=14. (F) Abrogation of the octopaminergic output of APL neurons via UAS-TBH-RNAi under APLGal4;Gal80ts at 30°C, compared to control animals carrying, but not expressing the transgene (UN). Dishabituation with a 45V electric shock (magenta) and YO (orange), although partially reduced, was sufficient in both conditions. n= 12. (G) Photoactivation of GABAergic neurons in OCT-habituated flies restored OCT avoidance only at the lower light intensity. n=20. (H) Photoactivation of GABAergic neurons in naïve flies resulted in light intensity dependent decrement in OCT avoidance. n=20. (I) Photoactivation of MBONs 8 and 9 under the MB110CGal4 driver in OCT-habituated flies restored OCT avoidance at the higher light intensities, but not at 4.5V (p=0.0062). n=14.

Ellipsoid Body Neurons do not participate in olfactory Habituation nor in Dishabituation.

Mean Performance Indices ± SEM are shown in all figures. The naïve response is shown in white bars and is always significantly different from the habituated response (grey bars). Heterosensory (footshock) dishabituation is indicated by the bolt and homosensory (YO), by the plume. Exposure to blue light is indicated by the LED bulbs. Stars in all Figures indicate significant differences (p<0.005), from the habituated response and daggers significant difference (p<0.005) from naïve avoidance. All statistical details are presented in Supplemental Table 1. (A) Schematic of Ellipsoid Body (EB) neurons. Created with BioRender.com/18aa3tx. (B) Functional silencing of the EB neurons with UAS-shibirets under RDLGal4 compared to control animals carrying, but not expressing the transgene (UN). Olfactory habituation and dishabituation with a 45V electric shock (magenta) and YO (orange), remained unaffected in both conditions. n=14. (C) Photoactivation of EB neurons under RDLGal4 in OCT-habituated flies with 3 light intensities did not restore avoidance, while a single footshock and YO exposure did. n=18. Additional controls run in parallel are presented in Supplemental Table 1. (D) Photoactivation of EB neurons in naïve flies with 3 light intensities did not affect OCT avoidance. n=12.

A model of the neuronal circuits underlying (A) Habituation Latency (B) Habituation and (C) Dishabituation.

Main neuronal compartments mediating the behavior comprise the antennal lobe (AL-green), the lateral horn (LH-orange) and the mushroom body (MB), divided in three lobes (α/β-pink, α’/β’-purple, γ-blue). In the stick figure the same neuronal compartments are presented as grey boxes oriented by a dashed line. Model created with BioRender.com/18aa3tx Distinct neuronal subsets are marked with different colors; Olfactory sensory neurons (OSNs-black), excitatory local interneurons (eLN-light orange), inhibitory local Interneurons (iLN-red), excitatory projection neurons (ePN-dark cyan/petrol), inhibitory local Interneurons (iLN-brown), mushroom body output neurons (MBONs-black), anterior paired lateral neuron (APL-black), PAM dopaminergic neurons (PAM-dark green), PPL1 dopaminergic neuronal subsets (PPL1-γ2α’1 & PPL1-γ1pedc both in yellow). Neurons not involved in the progression from latency to habituation and ultimately in the reinstatement of avoidance levels (dishabituation), are colored light grey for clarity. Synapses thought to be electrical connecting interneurons of the AL such as the eLN,iLN & ePN are indicated by -||-. Synapses and arrows show the projections of particular neurons (information path) in the circuit. Arrowheads indicate activation, while blunt arrows inhibition. A) Prolonged exposure to OCT negates the ePN inhibition, activating iPNs and drive habituation. B) To reinstate avoidance upon the presentation of a novel stimulus, neurotransmission from different neurons innervating the MB is required. For heterosensory dishabituation input from PPL1-γ2α’1 and PPL1-γ1pedc DANs to the MB is essential to encode the dishabituator (shock). In contrast, homosensory dishabituation requires synergy between the PAM cluster and APL neurons. Apl seems crucial for homosensory dishabituation potentially helping to differentiate the identity of the novel odor from the one habituated to. MB activation via PPL1, PAM and APL neuron switches again the MB output reactivating the α/β along with γ MBns, reinstating avoidance levels through LH activation, via different MBONs combinations.