Distinct Mechanisms for Synchronization and Temporal Patterning of Odor-Encoding Neural Assemblies

See allHide authors and affiliations

Science  08 Nov 1996:
Vol. 274, Issue 5289, pp. 976-979
DOI: 10.1126/science.274.5289.976


  • Fig. 1.

    Local neurons inhibit projection neurons monosynaptically. (A) Camera lucida drawing of a local neuron stained intracellularly from a dendrite by injection of cobalt hexamine. Dense, extensive arborizations occupy the entire neuropil, and no axon is present. Bar, 100 μm. (B) Simultaneous intracellular recordings from an antennal lobe local neuron (LN) and a postsynaptic projection neuron (PN), as well as the local field potential (LFP) from the ipsilateral mushroom body, during a response to a cherry odor in vivo. PN spikes are clipped. Calibrations: horizontal, 40 ms; vertical, 200 μV (LFP), 4 mV (PN), 5 mV (LN). (C) Direct current injection pulse (between arrowheads, 600 pA) in the LN [same pair as in (B)] inhibits tonic firing of PN (held depolarized by 300-pA constant current to evoke tonic spiking). The experiment was carried out in the absence of odor. The LN does not oscillate intrinsically upon depolarization. PN spikes are clipped. Calibrations: horizontal, 0.7 s; vertical, 32 mV (PN), 18 mV (LN).

  • Fig. 2.

    Picrotoxin injection in the antennal lobe selectively abolishes the oscillatory synchronization but not the responsiveness of PNs. [(A) to (D)] (top trace) LFP from mushroom body; (middle) simultaneous intracellular recording from antennal lobe PN during odor puff (indicated by horizontal bar); (bottom) sliding cross-correlation between LFP and PN traces (14). Only the relative (and not the absolute) values of the cross-correlation functions matter. High values in hot colors, low values in cold colors. (A) Control response to mint odor. The biphasic PN response and the prominent IPSPs are apparent during the odor response. The oscillatory LFP indicates synchronized and rhythmic firing of many other PNs during the odor response. The cross-correlation between PN and LFP shows a striped pattern with ∼50-ms period during the first half of the odor puff (vertical bar between 0.5 and 1.5 s), indicating an odor-evoked transient synchronization between this PN and the LFP (14). (B) Same pair as in (A), 2 min after pressure injection of 800 μM PCT into the antennal lobe. Although the response pattern of the PN to mint is not qualitatively altered (the periods of initial excitation, suppression, and subsequent excitation are preserved), the IPSPs have disappeared (indicating the block of LN-mediated fast inhibition), and the 20-Hz LFP oscillations are abolished. The cross-correlation is aperiodic, indicating desynchronization of the PN assembly representing mint. (C) Local injection of saline into the antennal lobe has no effect on synchronization and LFP oscillations, indicating that the manipulation per se does not disrupt the local circuits. A different animal was used from that in (A) and (B). (D) Local injection of PCT into the calyx of the mushroom body (where the axonal collaterals of PNs terminate) does not affect either the responsiveness of PNs to odors or their synchronization (assessed from the LFP oscillations or the periodic cross-correlation). [(A) to (D)] PN spikes are clipped. Calibrations (electrophysiological traces): horizontal, 1 s; vertical (in millivolts): LFP: [(A) and (B)] 0.2, (C) 0.5, (D) 0.3; PN: [(A) and (B)] 10, (C) 5, (D) 30.

  • Fig. 3.

    Effect of PCT applied simultaneously to the antennal lobe and mushroom body on synchronization and LFP oscillations. (A) Power spectrum calculated from the LFP oscillation (example in inset) evoked by a cherry odor (horizontal bar) in an intact animal (n = 20 odor presentations). A narrow peak is seen at ∼24 Hz. (B) Superfusion of 1 mM PCT onto the brain of the same animal rapidly eliminates the LFP oscillations, as a result of the desynchronization of PNs in the antennal lobe (Fig. 2). (C) Seven minutes later, odor puffs now evoke bursts of large-amplitude population spikes in the mushroom body (inset), whose power spectrum also shows a peak at 24 Hz, despite the desynchronization of PNs. Calibrations (insets): horizontal, 1 s; vertical, 500 μV.

  • Fig. 4.

    PCT injected in the antennal lobe (A) (1 mM) or perfused on the brain (B) (5 mM) abolishes PN synchronization but does not affect odor-specific temporal response patterns of PNs. Compare the slow response patterns of two different PNs (A and B) recorded intracellularly in vivo before (i and iii) and after (ii and iv) PCT application. Although PCT abolishes the LFP oscillations (caused by synchronized PN activity) and the trains of rhythmic IPSPs visible in individual PNs [arrowheads in (i)], the timing and duration of firing of the PNs are not altered, even when periods of inhibition appear to be caused by the PCT-sensitive IPSPs. (iii and iv) Post-stimulus time histograms (instantaneous frequency in hertz) constructed from repeated presentations of the same odor before (iii) and after (iv) PCT application for these two neurons. Trial numbers: Aiii, 10; Aiv, 10; Biii, 15; Biv, 4. Bin size, 50 ms. (A) and (B) are from different animals. PN spikes are clipped. Calibrations: horizontal (i to iv), 1 s; vertical: 200 μV (LFP); 10 mV (PN).

Stay Connected to Science