Strengthening the Alert Behavior

1.  Putting the alert behavior on stimulus control; only when cued

• in a series of other behaviors
• from a sit
• from a down
• while I am standing
• while I am seated
• while I am reclining

2.  Repeating the above for another person; only when cued

• in a series of other behaviors
• from a sit
• from a down
• while standing
• while seated
• while reclining

3. Repeating the above with distractions

• with another dog present
• multiple behaviors for a single reward
• in other locations
• inside
• outside
• at the car
• at the beach
• from a distance
• when Rich is here
• when Lucy is arriving
• when someone is at the gate

Variations on a Theme

Fluency The ability of the dog to perform a behavior under any and all circumstances. When a behavior is said to be truly fluent, it a) happens every time it's cued; and b) is never offered when it isn't cued.

Rules of Stimulus Control: 
1. Behavior is not offered in absence of the cue 
2. Behaviors is offered every time cue is given 
3. Behavior isn't offered in response to a different cue 
4. No other behavior occurs in response to the cue 


More work on the alert behavior and stimulus control today.  This time, seated in the living room.

Lily's mat is in the room.

We are working on Poke as another cued behavior.  One of many.

Sit.  Down.  Settle.  Brace.  Poke.  Go Lie Down. Show Me.  Place.  Visit.

etc etc etc

Not "Poke/No Poke".  But Poke as one of many.  Cue discrimination.

A good session.  (And excellent Latency as well.)

Fluency and Stimulus Control

Fluency The ability of the dog to perform a behavior under any and all circumstances. When a behavior is said to be truly fluent, it a) happens every time it's cued; and b) is never offered when it isn't cued.

Rules of Stimulus Control: 
1. Behavior is not offered in absence of the cue 
2. Behaviors is offered every time cue is given 
3. Behavior isn't offered in response to a different cue 
4. No other behavior occurs in response to the cue 

Today we are working on a tricky area - fluency of the alert behavior while I am lying on/in my bed.  Setting the foundation that this behavior is only done on cue.  No scent involved.  

Lily has never performed the alert behavior, "Poke", in this environment and with each of us in different positions.

So we began with just cueing the behavior and the physical ability of Lily to do the behavior lying down as well.  Later I will move her to standing on the bed and standing beside the bed.  But today we are working on teaching Lily that she can nudge me with her nose while lying beside me.

She positioned herself beside me, her head resting between her paws.

When I first cued "Poke" I expected that she would nudge my torso, but she nudged my arm instead, firmly enough to move my arm.  That's alright.....I think.  But it caused me to pause.  And to think it through.

I know that I don't want her nudging me on the face.  I know that I don't want her nudging me on my feet or my legs.  The first I just don't like.  The second I don't know if that would transfer to nudging my legs while upright and/or moving.  More to think about.

But, for now, prone and on the bed and assuming sleep, arm, torso, shoulder would all be acceptable.

Using the ABC (Antecedent (cue), Behavior, Click) model, the sessions went as follows:

A, B, C  X10

Repeat X10

Repeat X10.

All 3 sessions she firmly nudged my extended arm.

A, B, pause, 2-fer, C
 A, B, C
  A, B, C
    A, B, pause, 2-fer, C
      A, B, pause, 2-fer, C
        Repeat to total of 10

Did that for 2 sessions.  Still nudging my extended arm.

Then we began training fluency - poke/no poke

B2 (head resting between paws), C
  B2, C
    B2, C
      B2, C
         A,B,C

Lily is learning on her own, that a nudge is a nudge, whether nudging an arm or my torso.  We did that a couple of times and called it a day.

Scent Discrimination

Work continues separately on Phase 2, discriminating the target scent and only the target scent and indicating and on Phase 3,  building strength of the alert behavior.  Building persistence into the alert behavior.  But also building fluency - she will only perform the behavior when cued.  (See more on this in Phase 3.)

In this session I review the alert behavior.  I review touching the dummy (empty) canister.  And I review touching the target canister (with scent) and using it as the cue for the alert.

But the point of this session is to begin teaching Lily to discriminate between a dummy canister which requires only a touch and the target canister (with scent) which triggers the alert behavior.  I am making the explanation very simple, clicking as soon as her nose touches the dummy and cueing the alert behavior after she touches the target (hot) canister.

Poke

Touch Dummy

Target Scent/Poke

Dummy/Target Plus Verbal Cue

Dummy Target: Touch, C/T
Scent Target: Touch, Cue Poke, C/T

Thinking Outloud

Lily has three "get help" seizure response behaviors.  Two were taught when she was young, one recently.  Each of these behaviors is unique and she performs them with alacrity and enthusiasm.  It always amazes me that she knows them so well, remembers them so well, and performs them so well.

Lily's Emergency Call Box cue is my body unresponsive in a prone position.  Body position.  And at home or in the yard or garage.
Lily's Public Seizure Response cue is similar to the ECB behavior because it is cued by my unresponsive body, but the behavior is different.
Lily's Emergency Phone Retrieve is cued with a verbal cue.  To retrieve the dedicated phone.  When cued only.  The phone is to be ignored at all other times.

The seizure alert behavior must be performed the same way.  Only when cued, no matter what that cue is.

Her alert behavior is a strong, persistent and repetitive poke or prod to the center of my body.  The cue will be the scent of seizure.  For the moment, we are using the scent of Earl Grey tea to teach the chain.  Eventually we will change the cue from Earl Grey to seizure scent.

So. 

How to teach strength and persistence into the "poke".  In clicker training we want the dog to offer behaviors.  I am thinking of teaching this chain the way I have taught the response behaviors: a single high value treat reserved for this behavior, always "double cue" and always treat, working on the presumption that when it does happen, the pattern will be established and she will always be expecting a treat.

I already know that I won't be able to hide scent on my body when in a prone position on the floor, because that will trigger a different response.  But I can train lying prone on the bed and on each sofa in my house.  (and seated, standing, moving, inside and outside, at home and in public)

"Poke" only when cued.  I can cue and wait for two-fers or three-fers to build both strength and persistence.  And ignore voluntary pokes.  Or do a series and C/T for not poking.  I'm thinking this through very carefully, now that she knows the behavior and knows the verbal cue.

In teaching the scent as the cue I present the scent to Lily in some form - container or in my hand or with scent rubbed into my hand.  I am beginning with the scent in hand in a film canister.

Now I want to teach her that the canister itself is not a cue - it is the scent emanating from the canister.  This really is the same concept as the scent wheel.  So I am going to begin presenting both hot and dummy canisters during a training session.  Identical to look at.  But one containing scent and one not.

And.....I will utilize the "change the cue" formula, adapted to this situation.

D, touch, C/T
D, touch, C/T
D, touch, C/T
SS, touch, "Poke", C/T

OR should I reverse the ratio?

SS,touch, "Poke", C/T
SS, touch, "Poke", C/T
SS, touch, "Poke", C/T
D, touch, C/T

Still thinking this through before I begin.  Because I am thinking how this will impact really long term when I once again use the "change the cue" formula to transfer this training with a tea bag and substitute seizure scent.

Working for Green Beans

Lily's trying to lose a few pounds, so we are mixing kibble and green beans.  She brought me the clicker (her way of asking to play the clicker game) and was happy to work for veggies.

I'm now getting up to 6 seconds of duration on the touch of the film canister at the end of the scent stick, whether placed against something, on something, under something or just out at the open.  High or low.  Left or right.

But where I discovered today that I can't get duration on the canister - in fact I can barely get a touch at all - is when I place the canister on my cheek or on my upper arm or the top of my head.

So this is where we will be working next.

We have taken Lily's alert behavior outside:  seated, standing, lying down, moving and turning.  Off leash and on.  Left side and right side.  She still is tending to swing out of good Side position when nudging on the right side, so that is an area that needs some more work.  If I wait for two-fers I get them and the second nudge is always more forceful and demanding ... just what I want.

Even though I have withdrawn Lily from Phase 1 of the Seizure Alert Project, I want to complete her training of the nudge alerting behavior of all other Phases, using other scents as target scents.

Chaining Alert Behavior w/Sample: 3

"I would contaminate EVERYTHING with the treat scent--the apparatus, random spots on the wall, my shoes, you name it. Rather than get into a futile effort to maintain sterile procedures, I'd let the dog learn that the only odor that pays is the target odor."

~Steve White

Lily is such a good teacher!  And I learned so much during this morning's training sessions.

Day 3

My training goal was to "share the scent" by contaminating both hands that touched the scent canister with scent smell and residue.  The subject of contamination has come up several times in our planning conversations and Steve's quote above seems to me to be the most practical and "real life".  So I want to be teaching Lily early on that the only odor that pays is the target odor.

After testing the voracity of the Day 6 scent and determining that it could still be distinguished by Lily, I moved today's training into a new room.

The only change was to be delivering the treat from both hands (necessitating in holding the treats in both hands during the course of the morning's sessions AND in holding and presenting the sample canister from both hands.

AND adding movement to the exercise.  One part of the behavior selection criteria was that the alert behavior could be done while I was moving.  This is one part of the criteria that I have not trained as yet.  TRAINER ERROR:  Before beginning the behavior chain, I should have taught "Poke" while moving.

Don't worry.  This is only the first of several eye openers during this very important morning training.

Session 1

Treats and clicker held in right hand.

Session 2 is a repeat of Session1, except now the treats and clicker are held in my left hand.  That means that treat odor and residue from Session 1/right hand are now being transferred to the sample canister.  Continuing with movement left and right but in a very limited space.  Then, I commit my second TRAINER ERROR: I ask for the behavior with Lily in Tuck position.  Before including the chain in Tuck position, I should have first taught the alert behavior to Lily in Tuck position.  Watching closely, Lily tells me quite plainly that she understands "Tuck" but doesn't know how to "Poke" from there.  To her, it's an entirely new behavior.

Treats and clicker held in left hand.

In Sessions 3 and 4, I am seated.  Lily is familiar with the alert behavior when I am seated AND the behavior chain when I am seated.  The only difference in these two sessions is the sample delivery hand and the treat delivery hand.  More work here on understanding that the only odor that pays is the target odor.

Treats and clicker held in right hand.

Treats and clicker held in left hand.

In the fifth and final session of the morning, we move to another area of the room and I am standing once again.  Again, working with limited movement, which Lily handles well.  TRAINER ERROR: This time, I ask for the chain in both the Tuck AND the Side positions.

Review, Remind, Reteach

~Steve White, again

via Training Levels: Steps to Success

and Sue Ailsby

PLAN:  We have maximized the usefulness of Sample 3.  So I have several options open to me.  Because I had been advised by Lily's vet to give her a 2 week break from the scent testing anyway, even though I have only 2 original scent samples left in the freezer (PS), good training sense tells me to stop working the chain and go back and re-teach the alert behavior by itself from Tuck and Side  positions.  And, while I'm at it, review having Lily move from a down position to nose target my hand. That, too, will soon be a necessary part of the behavior chain.

By the time the 2 weeks is over, we should have a stronger basic foundation and can have built it back into the chain left side, right side, moving, tuck, side AND from lying down.....all ready to re-introduce the scent trigger.

UPDATE:  I love having a Levels trained dog!  This afternoon we had 4 quick, short sessions - two in Tuck position, two in Side position.  The first session on each side, I asked Lily to nose touch the back of my hand, placed at Lily's nose height, my hip height.  A handful of treats - perhaps 12 - 14.  The second session on each side found her volunteering the nose target in the proper place, the proper position.

No Flip.  Just reporting.  As Sue would say, "The Levels show you the holes."  We're taking care of it!

COMEBEFORES: I'm using an adaptation of Lazy Leash Level 2 Step 2, off leash, and doodling to step, cue touch, step, cue touch, turn, cue touch, etc both in Tuck position and in Side position.  Then I plan on repeating (reteaching) with the leash on.

Chaining Alert Behavior w/Sample: 2

Day 2 with Sample

• Sample description
• Origin 1: Mary McNeight
• Sample 3: Day 5
• hydrated; refrigerated
• film canister

Session 1

Introducing environmental distractions

training in front of an open door.

There was a brisk breeze blowing in the door.

Session 2

Same location, with the door closed.

Changing position in relation to the dog.

Session 3

Changing location.  Changing position.

Session 4

Session 5

Lily's poke is not as determined when I am seated,

so I begin 'asking' for two-fers.

Then I get the forceful nose touch that I am after.

NOTES:  I am glad that I modified my original alert behavior to the "forceful, persistent nose touch".  The initial "tug clothing" behavior is just about totally extinguished.  Only occasionally when asking for more than one touch to build force or persistence, does Lily take the clothing into her mouth.  For this reason, I will only be using two-fers when seated and only ask for two. The alert behavior continues to meet all of my initial criteria.  Keeping it simple (and effective) is making the explanation easy for Lily to understand.

Chaining Alert Behavior w/Sample: 1

Day 1 with Sample

• Sample description
• Origin 1: Mary McNeight
• Sample 3: Day 4
• hydrated; refrigerated
• film canister

Session 1

• Standing
• Present canister; cue Touch
• Move canister; cue Poke
• X12

Session 2

• Standing
• Present canister; cue Touch
• Move canister; no cue Poke
• Drop cue Touch
• Touch/Poke
• X13

Session 3

• Seated
• Present canister; cue Touch
• Move canister; cue Poke
• X12

Session 4

• Seated
• Present canister
• Move canister
• Get both Touch and Poke
• X13

Session 5

• Standing
• Different location in room
• Present canister
• Move canister
• Get both Touch and Poke
• X10

Comeafters

Review, Remind, Reteach

The alert behavior.  Drop the cue.  Reteach in different positions, different circumstances, different locations.  Getting the behavior exactly the way that I want it.

Outdoors, seated

Outdoors, seated, chaining

Outdoors, partially reclined

Outdoors, reclined

Outdoors, reclined, "twofers" to build strength, persistence

We will continue to generalize this behavior in all positions and many locations with the exception of me lying prone on the floor, which is Lily's cue to activate the Emergency Call Box.

Once the entire Alert Chain is joined, then the scent will become the cue for the Alerting behavior.

We will continue adding Distance and Distractions.

Beginning the Backchain

Comebefores:

• Level 1 Target
• nose touch to hand
• Alert Behavior:  Poke
• standing, seated, reclining
• both sides of body

Review the alert behavior

Begin building the chain.  

• Introducing Touch followed by Poke
• Left hand Touch
• Standing

Begin building the chain

• Touch followed by Poke
• Right hand Touch
• Standing

Understanding the chain

• presentation of hand triggers
• uncued alert behavior followed by
• reward

The Poke Behavior

Comebefores:  (Training Levels: Steps to Success)

• Level 1 Target

Standing:

Seated:

Reclining:

Determining the Behavior

Criteria:

• unique
• not a behavior she does naturally that could be misinterpreted
• not like any other trained behavior that she performs
• can be performed if I am standing, seated, lying down (in any position), moving
• unobtrusive and silent
• easily performed

Examples of Chaining Alerts

Steve White

Scott Smith