Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2016 Feb 23;113(8):E1089-97.
doi: 10.1073/pnas.1516134113. Epub 2016 Jan 19.

Cortical Cholinergic Signaling Controls the Detection of Cues

Affiliations
Free PMC article

Cortical Cholinergic Signaling Controls the Detection of Cues

Howard J Gritton et al. Proc Natl Acad Sci U S A. .
Free PMC article

Abstract

The cortical cholinergic input system has been described as a neuromodulator system that influences broadly defined behavioral and brain states. The discovery of phasic, trial-based increases in extracellular choline (transients), resulting from the hydrolysis of newly released acetylcholine (ACh), in the cortex of animals reporting the presence of cues suggests that ACh may have a more specialized role in cognitive processes. Here we expressed channelrhodopsin or halorhodopsin in basal forebrain cholinergic neurons of mice with optic fibers directed into this region and prefrontal cortex. Cholinergic transients, evoked in accordance with photostimulation parameters determined in vivo, were generated in mice performing a task necessitating the reporting of cue and noncue events. Generating cholinergic transients in conjunction with cues enhanced cue detection rates. Moreover, generating transients in noncued trials, where cholinergic transients normally are not observed, increased the number of invalid claims for cues. Enhancing hits and generating false alarms both scaled with stimulation intensity. Suppression of endogenous cholinergic activity during cued trials reduced hit rates. Cholinergic transients may be essential for synchronizing cortical neuronal output driven by salient cues and executing cue-guided responses.

Keywords: acetylcholine; attention; cortex; optogenetics.

Conflict of interest statement

The authors declare no conflict of interest.

Figures

Fig. 1.
Fig. 1.
Example of transfected cholinergic neurons in the basal forebrain expressing the reporter EYFP. The microphotographs show the middle slice of a confocal stack taken at the level of the ventral nucleus basalis of Meynert (coronal slice). Cholinergic neurons were visualized using an antibody against the vesicular acetylcholine transporter (VAChT; SI Materials and Methods; red; Upper Left). ChR2-H134R-EYFP–expressing neurons are in green (Lower Left). Merged microphotograph is on Right, with white arrows depicting VAChT+EYFP-immunopositive neurons and red arrows (on white contrast) depicting cholinergic neurons that were not transfected (10-μm scale inserted). Note that visualization of the colabeling of some neurons was outside this particular focal plane/slice but present in adjacent confocal slices. Neurons in the more ventral portion of this section were not transfected by the virus. The image represents the general finding that about two-thirds of cholinergic neurons in the nucleus basalis were transfected and that EYFP expression was restricted to cholinergic neurons (Figs. S1 and S3).
Fig. 2.
Fig. 2.
Prefrontal choline currents, recorded using choline-sensitive microelectrodes, as a function of laser stimulation power and duration. (A) Electrode configuration and placement in the prelimbic (Prl) cortex. Choline oxidase (ChOX) was immobilized on two of four ceramic-based platinum recording sites. (B) Changing the applied potential of 0.7 V, the optimum oxidation potential of the reporter molecule H2O2 (red: vs. the reference electrode) to 0.00 V (black), eliminated optogenetically evoked currents, confirming the cholinergic basis of currents and controlling for potential confounds resulting from laser stimulation. (C) Mean choline currents from all trials evoked by stimulation of ChR2-expressing cholinergic neurons in the basal forebrain (BF; 5–25 mW; 1,000 ms). (D) Increasing stimulation power resulted in higher transient amplitudes (post hoc multiple comparisons: *P < 0.05; **P < 0.01, ***P < 0.001). Amplitudes did not vary by stimulation duration, and the two factors did not interact. (E) Compared with the shorter stimulation period, 1,000-ms stimulation generated transients peaked later and required more time to return within 50% of baseline (see also Fig. S5; for cortically evoked currents, see Fig. S4; for the impact of cortical state on choline currents, see Fig. S6).
Fig. S1.
Fig. S1.
(A and B) Schematic illustration of the main groups of cortically projecting cholinergic cell groups (pink regions) in the anterior (A) and posterior (B) basal forebrain (al, ansa lenticularis; HDB, horizontal nucleus of the diagonal band; nbm, nucleus basalis of Meynert; SI, substantia innominata) (7). Largest (C, E, and G) and smallest (D, F, and H) extension of EYFP expression in the BF of ChAT-Cre mice infused with ChR2(H134R)-EYFP in the anterior (C and D) and posterior (E and F) BF and in cingulate cortex (G and H). (Scale bar, 10 µM.) The spots of signal in thalamic regions in C and E are artifacts that fluoresced at a wide range of excitation wavelengths and were not associated with neurons or neuron terminals. Compared with EYFP labeling in the BF of mouse 5B, the BF of mouse H9 showed relatively lower levels of EYFP label in the dorsal part of the anterior nBM and the posterior end of the nbM. However, expression levels of EYFP in cortical projection fields did not differ between these two mice, owing to overlapping projections from BF regions to medial prefrontal regions (4, 6). I shows the individual false alarm rates resulting from ChR2 stimulation in noncued trials in these two mice. Although mouse H9, the mouse with the relatively smallest extent of BF EYFP expression, generated relatively more false alarms at lower stimulation power, both mice performed ∼60% false alarms at higher stimulation power. The relatively high volume of viral construct infusions (Materials and Methods) ensured that BF EYFP expression among mice infused with the three constructs reliably included the main cholinergic cell groups.
Fig. S2.
Fig. S2.
Optic fiber placements for all animals used in this study. Fiber placements in mice used to demonstrate ChR2-evoked choline currents (A), ChR2 stimulation-evoked increases in hits and false alarms (C), Halo stimulation-induced decreases in hits (E), and effects of laser stimulation in mice expressing EYFP only (G). Placements for each group are summarized in B, D, F, and H. Optic fiber tips were placed at the transition between nbM and SI, and tips were confined within the following coordinates: AP, −0.34 to −0.82 mm; L, 1.5–2.15; V (from dura), 4.2–5.2 mm.
Fig. S3.
Fig. S3.
Largest (A) and smallest (B) extension of EYFP expression in the BF of ChAT-Cre mice infused with ChR2(H134R)-EYFP used in the anesthetized recording experiments. EYFP expression in mice 6–10 involved the posterior nbM and dorsal SI (B), whereas labeling in the BF of mice 9–19 extended into the anterior nbM, SI, and HDB (A). (Scale bar in A and B, 100 μm.) Optic fiber placements are shown in C and D. ANOVA (SI Results and Discussion) and post hoc comparisons indicated that choline amplitudes (E) at the two lower stimulation intensities were higher in the mouse with the smaller BF transfection space, whereas higher stimulation intensities yielded statistically similar choline amplitudes across both animals (variances based on the analyses of 10 traces per power level and mouse).
Fig. S4.
Fig. S4.
Choline currents evoked by photostimulation in prefrontal cortex. The tip of the optic fiber was placed in between the upper and lower pairs of the platinum recording sites and 0.5 mm away from the electrode. Photostimulation (5–25 mW; 1,000 ms) produced sizable currents that were subtracted from the currents recorded via choline-sensitive recording sites. (A) Peak amplitudes of choline currents evoked by BF photostimulation and photostimulation in prefrontal cortex (BF data are reproduced from Fig. 2D; see also Fig. S5). Lower stimulation intensities generated current amplitudes that did not differ between the two locations. At higher stimulation levels, cortically evoked currents appeared to reach greater amplitudes, but this result likely was confounded by unequal photoelectric effects on the lower vs. higher pairs of recording sites. (B) Choline traces evoked by PFC photostimulation (mean values from all subjects), with BF-evoked currents (from Fig. 2C) shown in gray.
Fig. S5.
Fig. S5.
Optogenetically evoked choline currents (taken from Fig. 2C; in gray) and prefrontal choline current recorded in rats performing a cued appetitive response task (taken from figure 3A in ref. 18). Cue onset (red) and the onset and duration of photostimulation are indicated. Note that the previous data in Parikh et al. (18) were recorded at 2 Hz, whereas in the present study, currents were sampled at 20 Hz.
Fig. S6.
Fig. S6.
Identification of low- vs. high-frequency cortical states and associated choline currents evoked by BF ChR2 stimulation (5–25 mW; 1,000 ms). Local field potentials were recorded via the same platinum recording sites used for the detection of choline currents and states were identified as described in SI Materials and Methods. (A) Example LFP spectrogram showing a spontaneous cortical state transition from a sentinel channel in a urethane-anesthetized mouse used in this study. Periods with sparse power above 2 Hz (inactive) are denoted by the black bar above the spectrogram, whereas periods of relative high frequency power above 2 Hz (active) are denoted by the blue bar. (B) Normalized PSD from the two states shown in A and coded by color. Power was normalized to total power within the time windows shown. (C) Histogram of the power ratio from the periods denoted in A and B. Power ratio compares the relative contribution of power above and power below 2 Hz (SI Materials and Methods). The dashed line denotes the cutoff used to define active and inactive states and served as the basis for all subsequent analysis. Peak amplitudes (D), time to peak (E), and decay time (time for the current to decrease by 50% from peak amplitude; F) did not differ between choline currents evoked in association with low- vs. high-frequency states.
Fig. 3.
Fig. 3.
Timeline of major experimental events, task trial types, and baseline performance. (A) ChAT-Cre mice first acquired the SAT over 8–12 wk. Thereafter, they received bilateral infusions of one of the virus constructs into the BF (Upper Right). Seven days later, optic fibers were implanted into the BF and mPFC. Mice resumed task practice while tethered for 2–3 wk. The effects of optical stimulation across various stimulation intensities were tested in 8–10 sessions over the next 20–30 d with tethered nonstimulation days intermixed (Left). (B) The task consisted of a random order of cued and noncued trials. Following either event, two nose-poke devices extended into the chambers and were retracted upon a nose-poke or following 4 s. Hits and correct rejections were rewarded with water, whereas misses and false alarms were not (Right Inset; arrows in the inset and depicting nose-poke selection are color-matched; half of the mice were trained with the nose-poke direction rules reversed). Following an intertrial interval of 12 ± 3 s, the next cue or noncue event commenced. The photographic inserts show a cue presentation with a mouse orienting toward the intelligence panel while positioned at the water port (Left), a subsequent hit, a noncue event, and a subsequent correct rejection (Right). (C and D) Baseline SAT performance during tethering by groups of mice to be infused with one of the three virus constructs (n = 9 ChR2, n = 5 Halo, n = 5 EYFP). Mice detected cues in a cue duration-dependent manner (C) and they correctly rejected <75% of noncue events (D). Performance did not differ between the three groups (see Results for statistical analyses).
Fig. 4.
Fig. 4.
Optogenetic stimulation of cholinergic neurons during cued trials (n = 9 ChR2 mice). (A) The onset of the blue light coincided with cue onset and light was terminated 1,000 ms later. (B) Hit rates, averaged over cue durations, increased in response to BF stimulation of ChR2-expressing cholinergic neurons. (C) The effects of power significantly interacted with cue duration, reflecting significant increases in hits to shortest and medium-duration cues. Post hoc one-way ANOVAs indicated that by increasing power, and thus the amplitude of evoked release, stimulation resulted in increases in hits to shortest and medium-duration cues, but not to longest cues (post hoc comparisons: *P < 0.05; **P < 0.01). (D) ChR2 stimulation in mPFC did not significantly affect hit rates. (E) Neither BF stimulation (n = 3) nor mPFC stimulation (n = 5) affected the hit rates in EYFP-expressing control mice.
Fig. S7.
Fig. S7.
Response times for hits in trials with ChR2 stimulation of the BF (A) or PFC (B), or with photostimulation of the BF and PFC in mice expressing EYFP only (C and D). The effects of ChR2 stimulation power and cue duration on response times for hits interacted significantly (SI Results and Discussion). Post hoc ANOVAs and multiple comparisons (in A) indicated that the interaction mainly reflected that response times for hits to longest cues were relatively shortest at 16–20 mW.
Fig. 5.
Fig. 5.
Effects of optogenetic activation of cholinergic neurons on noncued trials (n = 9 ChR2 mice). (A) On noncued trials, laser stimulation began 1,000 ms before, and ended coincident with, extension of the nose-poke devices into the operant chamber. Increasing levels of stimulation power systematically enhanced false alarms when applied bilaterally to the BF (B) or unilaterally just to the right mPFC (C; post hoc comparisons: *P < 0.05; **P < 0.01, ***P,0.001). In control animals expressing EYFP, neither BF (D; n = 3) nor mPFC stimulation (E; n = 5) significantly affected the false alarm rates.
Fig. S8.
Fig. S8.
Response times for false alarms in trials with ChR2 stimulation of the BF (A) or PFC (B), or with photostimulation of the BF and PFC in mice expressing EYFP only (C and D). Response times for false alarms were not affected by photostimulation.
Fig. 6.
Fig. 6.
Suppression of cholinergic activity on cued trials. (A) The 589-nm laser was turned on 50 ms before the onset of the cue to fully suppress endogenous cholinergic signaling. (B) BF stimulation in Halo-expressing mice (n = 5) decreased hit rates, with increasing laser power producing greater effects. Although the effect of Halo stimulation appeared most robust for hits to longest cues, the interaction between cue duration and laser power did not reach statistical significance (post hoc comparisons of main effect of power: *P < 0.05). Halo BF stimulation neither affected the relative number of false alarms (C) nor omissions (D). Halo stimulation in the mPFC did not affect hit rates (E).
Fig. S9.
Fig. S9.
Response times for hits (A and B) and false alarms (C and D) in trials with Halo activation of the BF (Left) and PFC (Right). Response times were not significantly affected by Halo photoactivation.

Similar articles

See all similar articles

Cited by 51 articles

See all "Cited by" articles

Publication types

LinkOut - more resources

Feedback