I'd like to thank everyone who responded to my query about unoriented cells
in V1. Included below, with the permissions of the authors, are the messages
that I received, prefaced by a copy of my original query for context. The
respondents were (in alphabetical order):
Lisa Croner <croner@salk.edu>
David Grosof <grosof@am.seer.wustl.edu>
Moshe Gur <mogi@biomed.technion.ac.il>
Ehud Kaplan <kaplane@rockvax.rockefeller.edu>
Valentine Marcar <vmarcar@kispi.unizh.ch>
Mark McCourt <mccourt@plains.nodak.edu>
Dario Ringach <dario@hobson.cns.nyu.edu>
Jeffrey D. Schall <schalljd@ctrvax.vanderbilt.edu>
Christopher Tyler <cwt@skivs.ski.org>
Several people pointed to recent articles on the topic which provide some of
the detailed information that I sought. The authors of those articles also
found evidence that the properties of cells inside and outside CO blobs
change more gradually than what was reported in some of the other articles
that I cited originally. Thus, this topic seems more controversial than I
realized at first.
Thank you again,
Mike
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Original posting
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Date: Fri, 18 Apr 1997 03:04:53 -0400 (EDT)
From: Michael Hucka <hucka@eecs.umich.edu>
To: CVNET Server <cvnet@skivs.ski.org>
Subject: Seeking data on unoriented visual neurons in mammalian V1
Although much emphasis is placed on the spatially oriented visual neurons in
V1 and other visual areas, it is known that a certain percentage of cells are
unoriented [1-7]. These tend to occur most in the CO blob layers, but they
also seem to appear outside of the CO blob regions (for example, in layers 5,
2-3, and 4ca [3]). Unoriented cortical cells tend to have circular-surround
receptive fields, though some have no inhibitory surround [7], and most cells
are selective for wavelength, though some are not [1,5]. The latter were
termed "broadband" cells by Livingstone & Hubel [1]. Unoriented cells in V1
tend to be complex cells [3], and they tend to respond to lower spatial
frequencies than do the more common oriented simple and complex cells [6].
I'm working on a simulation into which I would like to incorporate a model of
the broadband unoriented cortical cells. But I'm having a difficult time
finding data on the parameters of these cells. Can anyone suggest where I
can find the following kind of information?
1) What are the spatial-frequency bandwidths of cells in this class?
(Preferrably drawn from monkey data, but I'll take anything :-).)
2) Assuming these cells are appropriately modeled with difference-of-Gaussian
circular receptive fields of the form
k_center * G(sigma_center) - k_surround * G(sigma_surround)
what are appropriate mean values for the sigma's of the Gaussians and the
weighting constants k_center and k_surround?
3) Are there any theories about the functions of the broadband unoriented
cells?
4) Are there any existing models of the receptive fields of the broadband
unoriented cells?
I'd be grateful for any help in this area. -- Mike Hucka (hucka@umich.edu)
References:
[1] Livingstone, M.S., & Hubel, D.H. (1984). Anatomy and Physiology of a Color
System in the Primate Visual Cortex. Journal of Neuroscience,
4(1):309-356.
[2] De Valois, R.L., & De Valois, K.K. (1988). Spatial Vision. Oxford
University Press.
[3] DeBruyn, E.J., Casagrande, V.A., Beck, P.D., & Bonds, A.B. (1993).
Visual Resolution and Sensitivity of Single Cells in the Primary Visual
Cortex (V1) of a Nocturnal Primate (Bush Baby): Correlations with
Cortical Layers and Cytochrome Oxidase Patterns. Journal of
Neurophysiology, 69(1):3-18.
[4] Saito, H.-A., Tanaka, K., Fukada, Y., & Oyamada, H. (1988). Analysis of
Discontinuity in Visual Contours in Area 19 of the Cat. Journal of
Neuroscience, 8(4):1131-1143.
[5] Ts'o, D.Y., & Gilbert, C.D. (1988). The Organization of Chromatic and
Spatial Interactions in the Primate Striate Cortex. Journal of
Neuroscience, 8(5):1712-1727.
[6] Tootell, R.B.H., Silverman, M.S., Hamilton, S.L., Switkes, E., & De
Valois, R.L. (1988). Functional Anatomy of Macaque Striate Cortex. V.
Spatial Frequency. Journal of Neuroscience, 8(5):1610-1624.
[7] Young, R.A., & Lesperance, R.M. (1993). A Physiological Model of Motion
Analysis for Machine Vision. Technical Report GMR-7878, General Motors
Research Laboratories, Warren, Michigan.
-- Mike Hucka hucka@umich.edu http://www.eecs.umich.edu/~hucka University PhD to be, computational models of human visual processing (AI Lab) of UNIX systems administrator & programmer/analyst (EECS DCO) Michigan~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Replies (in alphabetical order) ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
From: Lisa Croner <croner@salk.edu> Date: Mon, 21 Apr 1997 10:48:31 -0700 (PDT)
Try this paper: Edwards, Purpura, & Kaplan, 1995, Contrast sensitivity and spatial frequency response of primate cortical neurons in and around the cytochrome oxidase blobs, Vision Research, 35(11): 1501-1523.
Good luck, Lisa Croner
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ From: "Grosof, David 314-362-2384" <GROSOF@am.seer.wustl.edu> Date: Mon, 21 Apr 1997 16:40:40 -0600 (CST)
You might want to try to get quantitative data about unoriented V1 cells from Dario Ringach at NYU. dario@cns.nyu.edu There is a big difference between unoriented layer 4 cells and unoriented cells in other layers. Good luck, David Grosof
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ From: "Prof. Moshe Gur" <mogi@biomed.technion.ac.il> Date: Wed, 23 Apr 1997 12:22:16 +0300
You should be aware that it is wrong to lump together data from cat and monkey or from different cortical layers so be careful. As to data, our findings from the awake monkey (eg, Snodderly & Gur, J. Neurosci, 1995) indicate that the only layer where non oriented (NO) cells are found is 4Cb; most are color opponent cells. In layer 4Ca all RFs are definitely oriented but not as sharply as in other layers. We have found that RFs in 4Cb are larger than RFs in other layers and most don't show significant surround inhibition, thus they would exhibit low pass characteristics. We assume that NO cells which are spontaneously active and very easy to stimulate exert a non specific inhibitory influence on cells in other layers such that in interaction with excitatory inputs the parvo system can generate RFs with very selective properties. Your, Moshe Gur
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ From: "Ehud Kaplan" <kaplane@rockvax.rockefeller.edu> Date: Mon, 21 Apr 97 07:17:44 -0400
Without getting too deeply into various controversies, I suggest that you read the following:
1) Leventhal, A.G., Thompson, K.G., Liu, D., Zhou, Y. & Ault, S.J. (1995). Concomitant sensitivity to orientation, direction, and color of cells in layers 2, 3, and 4 of monkey striate cortex. J.Neurosci. 15:1808-1818.
2) Edwards, D.P., Purpura, K. & Kaplan, E. (1995). Contrast sensitivity and spatial frequency response of primate cortical neurons in and around the cytochrome oxidase blobs. Vision Res. 35:1501-1523.
3) Edwards, D.P. & Kaplan, E. (1992). How sharp is the orientation tuning of single cells in the cytochrome oxidase blobs of primate visual cortex? Invest.Ophthalmol.Visual Sci. (suppl) 33(4):1255(#2812).
4) Croner, L.J. & Kaplan, E. (1995). Receptive fields of P and M ganglion cells across the primate retina. Vision Res. 35:7-24.
I think you should find these helpful.
Dr. Ehud Kaplan Jules & Doris Stein Research-to Prevent-Blindness Professor Departments of Ophthalmology, Physiology & Biophysics Box 1183 The Mount Sinai School of Medicine One Gustave Levy Place New York, NY, 10029
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ From: Valentine MARCAR <vmarcar@kispi.unizh.ch> Date: Tue, 22 Apr 1997 14:07:11 +0000
Dear Dr. Hucka,
in reply to your query on the CV net on the receptive field of V1 neurons. I have done some work on V1 cells in the course of which we measured spatial composition of the RF and the end-stopping. We did not look at spatial frequncy selectivity in any systematic way. I would suggest you write to Dr. Steve Raiguel at the KU Leuven in Belguim and ask for deteils. His e-mail address is:
steve.raiguel@med.kuleuven.bc.be . The only model of V1 receptive fields I am aware of is by R. Maex, who wrote his PhD thesis on the subject. Write to him at KULeuven Labo. Neuro- en Psychofysiologie Campus Gasthuisberg Herestraat, 49 B-3000 Leuven Belgium, Europe
I am not sure of his e-mail addres but you can try
reihoud.maex@med.kuleuven.ac.be . If this address does not you will just have to send him a letter.
Best wishes
Val Marcar
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ From: Mark McCourt <mccourt@plains.nodak.edu> Date: Mon, 21 Apr 1997 13:32:40 -0500
Dear Dr. (to be) Hucka:
My colleague (Barbara Blakeslee) and I have recently published (Vision Research, in press) a computational account of brightness perception in a variety of brightness illusions (e.g., grating induction, classical brightness contrast, Hermann's grid) which is based on low-frequency non-oriented units of the kind you refer to. We used volume-balanced DOG units which formed an octave-interval array. The table below indicates the space-constants we used. These were not inspired by physiologically derived estimates, however, because there are few detailed accounts of the properties of these units. We would be happy to send you a preprint of the paper if you are interested.
-Mark McCourt
Mechanism Space Constant (deg) Center Surround _______________________________ 1 .047 .093 2 .094 .188 3 .188 .375 4 .375 .75 5 .75 1.5 6 1.5 3 7 3 6
Mark E. McCourt, Ph.D. Voice: (701) 231-8625 Associate Professor Fax: (701) 231-8426 Department of Psychology North Dakota State University Fargo, ND 58105-5075
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ From: "Dario Ringach" <dario@hobson.cns.nyu.edu> Date: Mon, 28 Apr 1997 08:54:13 -0400
Mike, Hi-
> 1) What are the spatial-frequency bandwidths or cut-off frequencies of cells > in this class?
It depends where you are looking in the retina. It is non-uniform and varies quite a lot. Some of them have lower spatial frequencies than the oriented cells at the same location, but others can actually respond to very high spatial frequencies as well...
> 2) Are there any existing models of the receptive fields of the broadband > unoriented cells? (E.g., are these cells more appropriately modeled as > single Gaussian blobs or as DOGs?)
They look more like center-surround DoGs than single blobs. Almost always they have a surround.
> 3) Assuming these cells are appropriately modeled with difference-of-Gaussian > (DOG) circular receptive fields of the form > > k_center * G(sigma_center) - k_surround * G(sigma_surround) > > what are appropriate mean values for the sigma's of the Gaussians and the > weighting constants k_center and k_surround?
We probably have enough data to get the numbers out... but it has not been processed yet, so I don't have numbers to give you. The best you can do is to look at Hawken and Parker for the moment...
Are you just trying to do texture segmentation or something else???
-- Dario
-- Dario Ringach | office: (212) 998-7614 lab: (212) 998-7613 Center for Neural Science | home: (212) 727-9346 fax: (212) 995-4011 4 Washington Place, Rm 809 | New York University | e-mail: dario@cns.nyu.edu New York, NY 10003 | http://www.cns.nyu.edu/home/dario
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ From: schalljd@ctrvax.Vanderbilt.Edu (Jeffrey D. Schall) Date: Mon, 21 Apr 1997 10:45:11 -0500 (CDT)
You will find much useful information in Leventhal, et al. (1995, J. Neurosci. 15:1808).
Good luck
Jeffrey D. Schall, Ph.D. Vanderbilt Vision Research Center Department of Psychology Vanderbilt University Nashville, TN 37240
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ From: cwt@skivs.ski.org (Christopher Tyler) Date: Tue, 22 Apr 97 13:39:51 PDT
I applaud your interest in circularly symmetric cells, which have been negelected for far too long, in my opinion. Lei Liu, Cliff Schor and I review the literature on them in our 1992 Vision Research paper on dichoptic summation. Apparaently, most CC cells are binocular (!), making them a good substrate for the summation effects we found in 'rivalrous' stimuli at low contrast.
There is a potential artifact in recording such cells, however, which is that cells are supposed to lose their orientation tuning as they get metabolically degraded. I don't know how general this effect is, but it has been suggested as the basis for the 60% figure given by Baizer, Robinson & Dow (1977), the most complete reference on the topic.
Christopher W. Tyler, Ph.D. Associate Director Smith-Kettlewell Eye Research Institute 2232 Webster Street, San Francisco, CA 94115 USA Ph. 415-561-1640 Fax 415-561-1610 Email: cwt@skivs.ski.org Home Page: http://www.ski.org/cwt.html
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ The end. ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~