Fig 1.
Schematic drawing of the dynamic DNMTP task.
The sequence of four lever presses constituting a DNMTP trial. The base lever location where start and end of delay presses occurred was randomly selected for each trial. The sample lever was randomly selected as 90° to the left or right of the base lever. Levers 90° to the left or right of the base were extended for the choice. Reinforcement, indicated by *, was delivered to the drinking spout immediately above the lever following sample and correct (non-matching to sample) presses.
Fig 2.
Properties of recorded action potentials.
(A) Peak to trough action potential width vs. log firing rate for cells classified with each of the most commonly observed response types. (B) Tetrode recordings and (C) interspike interval (ISI) histogram from a neuron with narrow width (125 μs). (D) Tetrode recordings and (E) ISI histogram for a neuron with a wide width (344 μs). Calibration marks are 50 μV (vertical) and 200 μs (horizontal) in B and D. Red lines indicate 1 ms ISI (C, E).
Fig 3.
Peri-event time histograms and raster plots for different response types.
Histograms show 99% confidence interval used to define event related activity in blue. Raster plots show event markers for start (green), sample (blue), delay (red), and choice (yellow) lever presses. Responses include: (A) preparatory responses relative to start and delay lever presses; (B) movement I, (C) movement II, (D) lever press excitation, and (E) base lever press responses relative to start, sample, delay, and choice lever presses; (F) lever press/ reinforcement, (G) reinforcement, (H) reinforcement anticipation, (I) error, (J) post-reinforcement responses relative to sample and correct and incorrect choice responses, (K) delay-related responses relative to sample, delay, and correct and incorrect choice responses, and (L) lever press suppression. Activity is plotted in seconds from lever presses along the abscissa.
Fig 4.
Timing of different response types during DNMTP trials.
The average (thick bar) and SEM (thin bar) duration of increased activity is plotted relative to start, sample, delay, and choice lever presses. Responses were measured based on when activity was above the 99% confidence interval (see Fig 3). These analyses were restricted to responses sufficiently robust to remain consistently outside this limit during periods of elevated activity. Results are plotted for preparatory (Prep, N = 25 for start and N = 19 for delay responses), movement I (Move 1, N = 55), movement II (Move 2, N = 12), lever press (Lev, N = 24), base lever press (Base Lev, N = 13), post-reinforcement (Post R, N = 8), reinforcement excitation (R Excit, N = 31), reinforcement anticipation (R Antic, N = 17), error (N = 4), and delay (N = 19). Preparatory and movement responses are plotted relative to the lever presses they preceded. Reinforcement, error, and delay responses are plotted relative to the sample and choice lever presses that they followed.
Fig 5.
Heat maps showing spatial distribution of activity for different response types.
Activity recorded for the entire session (up to 60 trials or 60 minutes) is plotted as spikes/s in each of the bins in the 70 x 70 array that met the minimum criteria of 3 visits and 0.2 s occupancy (see color scale to the left of each plot). Results are shown for: (A) preparation, (B) movement I, (C) movement II, (D) lever press, (E) base lever press, (F) reinforcement excitation, (G) reinforcement anticipation, (H) error, (I) post-reinforcement, (J) delay, and (K) reinforcement suppression. Maps are oriented with panels containing levers and drinking spouts numbered 1 to 4 centered in the corners moving clockwise from the upper left (L). The locations of base levers (either two opposite or all four) for the session depicted are indicated by small white arrows.
Fig 6.
Variability in spatial heat maps, raster-plots, and peri-event time histograms (PETH) for neurons with movement I responses.
PETHs and raster plots are aligned with delay lever presses to show increased activity before and after this response. The 99% confidence is indicated on PETHs in blue. Raster plots show event markers for sample (blue), delay (red), and choice (yellow) lever presses. Heat maps are oriented as in Fig 5, with activity in spikes/s indicated on the scale to the left of each plot and locations of base levers marked with small white arrows.
Fig 7.
Spatial heat maps (A to F) and peri-event time histograms (PETH) and raster plots (G to L) for neurons with spatially-restricted responses.
Heat maps, oriented as in Fig 5, with activity in spikes/s indicated on the scale to the left of each plot and locations of base levers marked with small white arrows. PETHs show average spikes/s relative to different lever press responses. The 99% confidence intervals are indicated by the blue areas in each PETH. Raster plots are aligned with PETHs and show event markers for start (green), sample (blue), delay (red), and choice (yellow) responses for individual trials. Results are shown for delay (A, B, F, G, H, L), base lever press (C, I), reinforcement excitation (D, J), and reinforcement anticipation (E, K) responses.
Fig 8.
Directional coding and timing properties of movement II neurons.
Raster plots and PETHs (aligned with sample and choice responses) reveal characteristic pattern of elevated activity more prolonged prior to sample than choice for two neurons (A and B). Event markers in raster plots show start (green), sample (blue), delay (red), and choice (yellow) responses in individual trials. Spatial heat maps are shown based on the entire recording session (overall) and the 4 s period preceding sample, correct choice, and incorrect choice responses. DNMTP was trained with two base lever locations for both these examples (base lever locations indicated by white arrows in corners of overall maps). Activity is plotted as spikes/ s as indicated on the scale to the left of each heat map. The paths traveled by rats during the 4.0 s preceding sample and correct choice responses are shown in the panels to the right. Both examples show activity elevated selectively to the left of the base lever location for the overall and sample plots even though the paths indicate comparable travel to the right. Activity and path maps indicate that movement II cells were active while rats visited both base and sample/choice lever locations during the 4.0 s before sample and choice responses.
Fig 9.
Broadly distributed and spatially-restricted firing patterns for neurons without criterion event-related responses.
Spatial heat maps show the distribution of activity of neurons with broadly distributed (A to D) and restricted (E to H) spatial firing fields. Maps are oriented and neural activity is plotted as spikes/s as indicated by the scale to the left of each plot as in Figs 5–8.
Fig 10.
Anatomical localization of neurons exhibiting different response types.
Coronal sections are modified from drawings 4.2, 3.7, 3.2, and 2.7 mm anterior to Bregma in [36]. Tetrode tracks are subdivided at intervals of 0.5 mm. Cortical fields are marked for frontal 2 (FR2), anterior cingulate (AC), prelimbic (PL), infralimbic (IL), and medial orbital (MO) areas [after 5]. Locations are plotted for neurons classed as preparatory (P), movement I (1), movement II (2), lever press (L), base lever press (B), post-reinforcement (X), reinforcement (R), reinforcement anticipation (A), error (E), and delay (D) response types. Each line represents cells recorded on two consecutive depths as electrodes were advanced ventrally. Photomicrographs show representative examples of two of the tetrode tracks (as indicated by arrows).