Receptive Field Bar Mapper parameters

In addition to the shared capture parameters.

No parameter on this page discards data

Every one of them is re-read from the accumulated trials whenever it changes, so they all stay editable during acquisition.

The exceptions are the two capture parameters — channels and pre_ms / post_ms — which do rebuild the accumulators and are locked while acquiring.

Capture — editor front panel

Parameter	Default	Range	What it does
channels	none	—	Which channels get a map. Everything downstream is linear in this count; the main cost lever. Changing it resizes and clears the accumulators.
pre_ms	500 ms	0 – 10000	Window before the trigger. Doubles as the baseline for the z-score: the spontaneous rate and its spread are both measured here (§2.4.2). Too short and the z-scores are noise-scaled.
post_ms	1000 ms	10 – 10000	Window after the trigger. This is what limits how far along its axis the bar is followed — see Making the window and the sweep agree.

Angle convention — ANALYSIS

Name	Label	Type	Default	Values
angle_zero	Zero at	categorical	Right	Right, Up, Left, Down
angle_sense	Angles turn	categorical	Counter-clockwise	Counter-clockwise, Clockwise

Where the stimulus program's zero angle points, and which way increasing angles turn.

Angles are stored exactly as typed and converted to a canonical form (0 = right, counter-clockwise) only where they are used, so changing the convention re-interprets the table rather than rewriting it.

A convention wrong by 180° produces a plausible and entirely wrong map

Check yours against the stimulus program. Fiorani et al.'s Appendix A, for instance, uses zero at left, counterclockwise.

Stimulus geometry — ANALYSIS

These describe the experiment. They must match what the stimulus program actually did; they are not free parameters.

Name	Label	Type	Default	Range	Step	What it does
speed_deg_per_sec	Speed	float, deg/s	10	0.1 – 200	0.1	Bar speed along its axis of motion. Sets the scale factor from time to space: one sample becomes speed / sample rate degrees. Getting it wrong scales the whole map, with nothing else looking odd.
sweep_start_deg	Sweep start	float, deg	-15	-180 – 180	0.1	Where the bar's centre was, along its own axis, at the trigger. Usually negative: the sweep starts off to one side and crosses the centre part-way through — for a sweep through the centre this is −travelDistance/2. Getting it wrong translates the whole map along each direction of motion, which shows up as a blurred or displaced receptive field.
latency_ms	Latency	float, ms	60	0 – 500	1	Neuronal latency, subtracted before time becomes space. Without it each direction's response is displaced along its own direction of motion and the combined field is inflated (their Fig. 3); too large a latency inflates it the other way. The right value maximises the map peak, which is what the latency scan exploits (§2.4.5).

How the response is read — ANALYSIS

Name	Label	Type	Default	Range	Step	What it does
smoothing_sigma_ms	Smoothing	float, ms	100	0 – 2000	1	Gaussian sigma applied to the z-scored trace, kernel truncated at 4σ (§2.4.3). Too narrow and high-frequency noise displaces the peak; too wide and the mapped field is inflated. A rule of thumb is a quarter of the time the bar takes to cross the expected receptive field. At 10 deg/s, 100 ms is 1 degree of travel.
use_absolute_z	Absolute z	bool	false	—	—	Rectifies the profile, so suppression counts as a response (§2.4.4). Turn it on to map an inhibitory receptive field. The map's sign then becomes meaningless, and the colour scale is relabelled \|z\|.
combine_mode	Combine	categorical	Arithmetic	Arithmetic, Geometric, Product	—	How the N directions are combined at each map pixel. See below.

Combine modes

Mode	What it is	When
Arithmetic	The mean	The paper's default, and the only mode its error figures were measured with
Geometric	The nth root of the product, with the sign restored	Sharper, but easily destroyed by one direction that responded near zero
Product	The plain product	Values are in zⁿ, not z, and the colour scale says so

The map itself — ANALYSIS

Name	Label	Type	Default	Range	Step	What it does
map_pixels	Map size	int, px	201	21 – 601	—	Map width and height. Forced odd internally so there is a true centre pixel. Cost is quadratic in this.
deg_per_pixel	Resolution	float, deg	0.1	0.01 – 2.0	0.01	Degrees of visual angle per map pixel. Together with Map size this fixes the extent: span = size × resolution, 20.1° square by default. Finer than the smoothing width buys nothing but pixels.
map_centre_x	Map centre X	float, deg	0	-90 – 90	0.1	Visual-field x of the map's centre pixel, in the same frame the sweep angles are in (+x right, +y up).
map_centre_y	Map centre Y	float, deg	0	-90 – 90	0.1	Visual-field y of the map's centre pixel. Move the map to where the receptive field is rather than enlarging it — cost is quadratic in size, and free in centre.
border_fraction	Border	float	0.76	0.1 – 0.99	0.01	Fraction of the map peak at which the receptive-field border is drawn, and therefore what the reported area, equivalent diameter and bounding box mean. Not 0.5 — 0.76 is the paper's empirical correction (§3.1.1) for the enlargement smoothing and back-projection introduce. Changing it changes the reported sizes, not the map.

Display units — ANALYSIS

Three parameters that change how numbers are written, never what they are. Degrees of visual angle stay the unit everything is computed, stored and exported in; these convert at the edges — the ANALYSIS fields and the map panels — so switching unit cannot move a receptive field or change a map by a single pixel. They are saved with the signal chain, because the rig does not change between sessions.

Name	Label	Type	Default	Range	Step	What it does
display_unit	Show units in	categorical	Degrees	Degrees, Millimetres, Screen pixels	—	The unit for Speed, Sweep start, Resolution, Map centre X/Y, the map axes and the RF readout. Everything else — latency, smoothing, border, map size — is a time, a ratio or a count of map pixels, and does not change.
viewing_distance_mm	Viewing distance	float, mm	570	10 – 5000	1	Eye to screen. Greyed out while the unit is Degrees.
screen_px_per_mm	Screen resolution	float, px/mm	3.6	0.1 – 100	0.1	Screen pixels per millimetre — 3.6 px/mm is about 91 ppi. Greyed out unless the unit is Screen pixels.

The conversion is the small-angle factor:

mm per degree = viewing distance × π / 180        (9.95 mm/deg at 570 mm)
px per degree = mm per degree × screen px per mm  (35.8 px/deg at 3.6 px/mm)

One scale for positions and extents alike, so a map pixel is the same size in millimetres wherever it sits. The exact conversion for a position is d · tan θ, which is smaller than the linear one — the linear factor under-reports position by about 4% at 20° eccentricity, and by about 1% at 10°. That is the price of a single number per field: the tangent form would convert the map centre and the resolution by different rules and leave the map grid non-uniform in millimetres. Map far into the periphery and read the positions in degrees.

Typing is converted the same way and clamped in degrees, so a parameter's range never changes with the unit: with the unit set to millimetres, a Speed typed as 2000 mm/s comes back as 1990 mm/s — 200 deg/s, the parameter's own maximum.

A viewing distance or pixel pitch of zero — a rig nobody has measured yet — falls back to degrees rather than showing every position as 0 mm.

Sweep angles — ANALYSIS → DIRECTIONS...

Not registered parameters: they are per-trigger-source state, persisted alongside the trigger table in the signal chain, in a trigger-settings file, and in a saved session.

Field	Default	Notes
Angle, per condition	unset	Direction of motion for that condition, in the convention above. A condition with no angle contributes nothing to the map — it is not treated as 0°.

The direction generator

Also per-plugin state rather than a registered parameter; saved with the signal chain and with the trigger table.

Field	Default	What it sets
Count	8	How many directions, evenly spaced around the circle
First trigger number	1	TTL line for the first condition, 1-based as the trigger table shows it
One line per direction	off	off: all directions armed on that one line, told apart by their messages; on: line, line+1, line+2, …
Arm message base	TRIALTYPE	Text before the number
First arm number	0	Number for the first direction; steps by one per direction
Arm message suffix	TIMESEQUENCE	Text after the number
First angle	0 deg	Angle of the first direction

For example, VSTIM: TRIALTYPE, 200 and TIMESEQUENCE generate VSTIM: TRIALTYPE 200 TIMESEQUENCE, … 201 …, and so on; the fields exist so the pattern can be matched to whatever your stimulus program sends.

The suffix is not decoration

TRIALTYPE 3 also contains-matches TRIALTYPE 30, and a trial-end message repeating the trial type would re-arm the source at trial end. Pick text that appears in the trial-start message and not in the trial-end one, and clear it only if your messages carry their own boundary.

The number steps up by one per direction whether or not the TTL line does. REPLACE discards the existing sources rather than appending.

Canvas display controls

Not registered parameters.

Control	Values	Default
View	Map, Traces	Map
Columns, Size	grid layout	—
POLAR	on / off	off
SAME SCALE	on / off	off — per-panel scaling is what makes a weak channel visible

Session arrays

Everything Triggered Average writes, plus:

Array	Shape	dtype	Contents
maps	(channels, pixels, pixels)	float32	The back-projection maps
map_estimates	(channels, 7)	float64	peak, centre_x_deg, centre_y_deg, area_pixels, equivalent_diameter_deg, width_deg, height_deg
map_valid	(channels,)	int32	Whether each channel's mapping is usable
map_channel_indices	(channels,)	int32	Global channel indices

Manifest attributes: map_pixels, map_degrees_per_pixel, map_centre_x_deg, map_centre_y_deg, and map_estimate_fields — the comma-separated column names of map_estimates.

See Saved sessions.